Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Hierarchical Interactive Architecture based on Coalition Formation for Neighborhood System Decision Making

Hierarchical Interactive Architecture based on Coalition Formation for Neighborhood System... In recent years, significant efforts for improving the technical and the economic performance of smart grids have been applied with the presence of different players making decisions in these grids. This paper proposes a bilevel hierarchical structure for designing and planning distributed energy resources in home microgrids (H-MGs). Such a small-scale grid inside the market environment permits energy exchange among distributed energy resources and home microgrids through a pool market. In this paper, each home microgrid’s planner maximizes the performance of distributed energy resources while exchanging energy with other H-MGs. The problem is formulated as a high-level problem decomposed into a set of low-level market clearing problems. The global optimal performance in terms of energy cost is met for a market structure (H-MGs, consumers, and retailers) at a Nash optimum point for the formulated scheduling game considering local and general constraints on the spot market. In general, the upper-level structure is based on H-MG generation competition for maximizing their individual and/or group income in the process of forming coalition with other H-MGs. In addition, the functionality of the lower-level of the hierarchical structure is governed by a market clearing based on the price response method by all the DERs enabling H-MGs to change the spot market strategic behavior. Prices are obtained as dual variables of power balance equations. This paper will investigate a set of bilevel games, including nonlinear programming problems solved through a complementary problem method. Using the binary theory, the bilevel hierarchical structure will be replanned using a nonlinear problem. Results prove that the proposed structure can increase the players’s profit. Keywords: Smart grids, electricity generation market, coalition formation, responsive load, bidding strategy, bilevel programming. WT wind turbine Nomenclature RET retailer Acronyms A, B, C home microgrids CHP combined heat and power Indices EB electrical boiler w the number of scenario (w 2 f1, 2, . . . ,Wg) EHP electrical heat pump i=m the number of H-MG (i 2 f1, 2, . . . ,Ig) ES+/ES- energy storage (ES) during charg- j the number of DERs existing in H-MG i (j 2 ing/discharging mode f1, 2, . . . ,Jg) GB gas boiler t time (t 2 f1, 2, . . . ,T g) k+/k- buying/ selling power from/to H-MG i/ Re- D the amount of electrical load consumer tailer k DERs’s demand (i.e., EB, EHP, ES) and/or MCP market clearing price consumers existing in H-MG i DR+, DR- amount of responsive load demand (RLD) k the number of a retailer (k 2 f1, 2, . . . ,ng) that goes from/comes to other time period e/h electrical/thermal to/from t FU fuel consumed by thermal DERs (i.e., CHP, STP solar thermal panel EB) TES thermal energy storage Email address: Jabervalinejad@vt.edu Corresponding author (Jaber Valinejad) Preprint submitted to International Journal of Electrical Power & Energy Systems October 17, 2019 arXiv:1910.07039v1 [math.OC] 15 Oct 2019 Constants 1. Introduction CHP,h the thermal efficiency of CHP ij CHP,e 0CHP,e Demand-side management (DSM) topics are focused on / the electrical efficiency of CHP ij ij energy consumption control at the consumer side [1, 2]. FU Such energy control is coordinated by electric utilities, com- the offer price of natural gas ES,ini panies and enterprises without controlling distributed energy SOC SOC initial and final value related to ES j in ij resources [3, 4]. When the latter is controlled, the topic ES,end H-MG i [kW] SOC ij would be defined as an energy management [5]. TES,ini SOC SOC initial and final value related to TES j in ij demandThis objective of this paper is to first propose a base TES,end H-MG i [kW] SOC framework for the demand of consumers encompassing H- ij EB TES ,  the thermal efficiency of EB/TES MGs, and second investigate profits that can be made from ij ij EHP COP coefficient of performance of EHP operating H-MGs independently or in a coalitional structure MC marginal cost of DER j in H-MG i [£/kWh] in a daily electricity market [6, 7]. For these purposes, en- ij ergy exchange among H-MGs is formulated as a scheduling P , P maximum electrical/thermal power gener- ij ij game. In this paper, competitive monopolies are modeled ated by DER j in H-MG i [kW] and simulated in a formulated convex optimization prob- Parameters MCP lem [8, 9]. These monopolies are based on three contra- the value of predicted electrical clearing tw dictory objectives: H-MGs’s and retailers’s income maximiza- price at time t under scenario w [£/kWh] D,e tion, consumers’s cost reduction, and demand peak reduction P predicted load power consumed by the con- itw [10]. sumers in H-MG i at time t under scenario w With the aim of a convex optimization problem, a bilevel [kW] hierarchical interactive architecture (BL-HIA) algorithm on Functions the condition of reaching a maximum profit is proposed for R the revenue of DER j related to H-MG i [£] ij both consumer and the power generator side [11, 12]. The R the revenue of H-MG i in independent oper- optimum performance problem is presented for all DERs ex- ation [£] isting in multiple H-MGs as a BL-HIA so that it is a mixed inte- im R the revenue resulting from coalitional oper- ger programming problem [13, 14]. The upper-level targets ation between H-MG i and H-MG m [£] maximizing H-MGs’s profit through energy exchange among Decision Variables H-MGs as well as H-MGs and retailers for a central optimal X,e Y,h / electrical/ thermal selling price bid by DER j ijtw ijtw performance of a decision maker [14, 15]. On the other under scenario w at time t [£/kWh] hand, the lower-level of the hierarchal structure of the prob- e h P /P electrical/thermal power generated by DER ijtw ijtw lem represents an equilibrium problem incorporating DSM j at time t under scenario w [kW] for an optimal performance of multiple H-MGs [1, 5]. This P electrical power sold to/bought from H-MG iktw way, a central optimum performance decision maker for en- i/ Retailer k at time t under scenario w [kW] ergy optimum exchange among H-MGs with each other and D,e P active power consumed by the consumers at itw with retailers in an independent and a coalitional perfor- H-MG i at time t under scenario w [kW] mance and with the aim of reaching H-MGs’s maximum profit 0D,e P shifted load power from time interval t to itt’w is included at an upper-level decision maker while consider- time interval t at H-MG i under scenario w ing an independent or a coalitional performance of all H-MGs [kW] [16–18]. The interaction between the two levels of the hier- EB,e P , consumed load power by EB/EHP j at H-MG ijtw archal structure of the game is a factor of searching for the op- EHP,e i at time t under scenario w [kW] timal solution at both levels [4, 5]. Considering the optimal ijtw MCP,e , electrical and thermal market clearing price scheduling of all H-MGs and DERs existing in them in Multi- MCP,h at time t [£/kWh] ple H-MGs requires solving mathematical program with equi- e- e+ P , P buying/selling electrical power from H-MG i librium constraints (MPEC) equivalent to bilevel problem. imtw imtw This bilevel problem can be looked at as a multiple-leader- to H-MG m [kW] e-=+ e-=+ common-follower game. The aim of implementing this game P /P buying/ selling electrical power from/to H- iktw kitw structure is finding a final equilibrium point in which none MG i/ Retailer k [kW] h- h+ of H-MGs and consumers can increase their profit by chang- P /P buying/selling thermal power from/to H- imtw imtw ing in the generation and the consumption schedules. Fur- MG i/H-MG m [kW] e- e+ thermore, the BL-HIA structure accounts for decisions result- / supply bids for buying/ selling electrical iktw iktw ing from forming a coalition among H-MGs to maximize the power from/to H-MG i/Retailer k [kW] h- h+ profit and also exchange energy among them. / supply bids for buying/selling thermal power imtw imtw The contribution in this work can be summarized as fol- from/to H-MG i/H-MG m [kW] lows: First, the proposed BL-HIA structure is preferable over the proposed structure in [19] as it is a multi-ownership structure that permits, forming coalition among H-MGs and D E R D ER D ER DER DER DER DER FL:A...SL:BC H- H- FL:B...SL:AC MG MG FL:C...SL:AB DE R FL:AB...SL:C FL:AC...SL:B FL:BC...SL:A D E R H- MG CEMS EMS DER:CHP-WT-STP-EB-EHP-ES- GB UL: upper level LL: lower leveL H-MG A and B :AB D ER DER DER H- H- MG MG D E R H- MG CEMS EMS DER:CHP-WT-STP-EB-EHP-ES- GB UL: upper level LL: lower leveL H-MG A and B :AB FL:A...SL:BC FL:B...SL:AC FL:C...SL:AB FL:AB...SL:C FL:AC...SL:B FL:BC...SL:A structure that is a decision-making problem including sev- variables DERs ,consumers , DER:CHP-WT-STP- load shifting of building with EB-EHP-ES-GB eral agents which try to optimize their corresponding objec- UL: upper level higher priority LL: lower level tive functions on a connectable dependent set. An agent is, AB:Coalition between DER H-MG A and B in fact, an object which can act as a DER or connected to DER profit profit other units. The BL-HIA structure is shown in Figure. 2. The maximization of the maximization of the upper-level problem states maximizing the profit of H-MGs profit of each H-MG profit of each H-MG Coalition with higher priority with higher priority having a higher priority of operating in an independent or a UL:A...LL:BC load shifting load shifting coalitional operation on the condition of: 1) satisfying upper- UL:B...LL:AC profit profit UL:C...LL:AB level constraints and 2) satisfying a set of lower-level prob- Upper- Lower- UL:AB...LL:C DERs variables ,consumers ,the H-MG ,load lems. H-MGs with a higher priority of operation in the upper- shifting with lower priority ,electricity prices UL:AC...LL:B level level and market hourly heat UL:BC...LL:A CEMS EMS level problem are identified based on their price bids. Upper- level constraints include limits on the quantity and supply bids of DERs resources, the minimum accessible power ca- Figure 1: The proposed BL-HIA structure illustrating a variety of coali- pacity by the market regulator, and buying/selling quantities tion formation among H-MGs. UL:A...LL:BC implies that the HMG A is modeled in the upper level while HMGs B and C are modeled in lower by H-MGs and retailers. Lower-level problem states the mar- level.Distributed energy resources encompass set of distributed generation, ket clearing prices (MCP) with the aim of maximizing the electrical and thermal energy storage. profit of H-MGs having lower priority of operation subject to meeting equilibrium constraints for each H-MG, genera- tion/consumption limits, and the number of consumers par- explicitly increasing the competition among H-MGs and con- ticipating in the DSM program. sumers rather than an independent operation of H-MGs. Sec- As it is observed in Figure. 2, the higher priority of H- ond, the BL-HIA is adequate for modelling problems with MGs operation (in an independent or a coalitional opera- several leaders (i.e., H-MGs) having their own individual ob- tion) is defined by the energy excess/shortage gap and sup- jective functions when operating independently or in a coali- ply/demand bids that permits a maximization of an expected tional manner (upper-level problem). Such a game is to opti- profit. Maximizing the profit of each H-MG is achieved by mize several followers (i.e., consumers inserted in the bilevel considering the fact that each agent at the lower-level prob- structure). These models are related to situations where ac- lem shows an optimal operation in correspondence to the in- tions and followers’s performance in BL-HIA have a signifi- come of H-MGs with higher priority of operation based on cant effect on decisions made by leaders. This fact is related the offered price. This optimal operation includes an estima- to the case in which H-MGs’s profit (a leader) depends on tion of the demand supplied and shifted by each consumer the amount of energy which is sold for supplying consumers in an independent or a coalitional operation. It must be em- existing in the power grid (as follower). The general view phasized that the competition among H-MGs with higher and of the hierarchical structure and optimization problems has lower priority (or competitor H-MGs) are explicitly modelled been shown with the proposed model in Figure. 2. Third, a at upper and lower-levels. It must be noted that upper-level better strategy maximizing consumers’s satisfaction in terms and lower-level problems shown in Figure. 2 become related of demand supply and H-MGs’s profit is presented in compar- to each other. In other words, lower-level problems estimate ison to a single level structure. Finally, the BL-HIA structure is the price and the quantity of competitor H-MGs which di- solved by formulating an equivalent one level mixed-integer rectly affect the profit of H-MGs inserted in the above prob- nonlinear programming (MINLP) problem deploying the KKT lem. In other words, decisions related to forming a coalition optimization conditions. The innovations in this paper can be and taking a bidding strategy by the competitor strategic H- summarized as follows: MGs in the upper-level problem have also a significant effect 1. An optimum programming solution within H-MGs gen- on the MCP resulted from the lower-level problem. eration as a BL-HIA structure; The BL-HIA structure is shown in Figure. 1. In BL-HIA 2. Providing a multiple-leader-common-follower game structure is also considered for modelling the uncertainty of which states the effectiveness of the market competition pool prices, electrical/thermal load demand, and H-MGs buy- in multiple H-MGs through solving a BL-HIA structure; ing/selling prices. The proposed bilevel model has been sim- 3. Developing a new model for demand-side management; plified as a one-level problem using KKT method [20] for the 4. Accommodating both DR resources and storage devices sake of convexification. in the market operation to achieve a comprehensive so- lution exploiting all flexibilities; and 3. Decision-Making Process by BL-HIA Structure 5. Proposing an advanced electricity market for active dis- tribution networks based on game theory; The decision-making process in the BL-HIA structure of H- MGs, consumers and retailers can be summarized as shown in Figure. 3. At the beginning of the scheduling horizon, each 2. The Proposed BL-HIA Structure H-MG presents the necessary decisions on DSM for an inde- The problem encountered by the H-MGs for an indepen- pendent or a coalitional operation with other H-MGs. More- dent or a coalitional operation can be modelled as a bilevel over, supply/demand bids are provided to the consumers dur- 3 Wind speed Solar irradiance 4.1. DR Objective Function and Constraints prediction prediction Power Power Electricity price generated by Load forcasting D,e DSM MCP MCP generated by PV prediction ˜ ˜ WT max R = 0.5 P   -  (1) i it’w tw t’w Modelling uncertainty using Taghochi method 0 t =1 Retailers H-MGs with higher priority in independent or The profit resulting from the participation of consumers in (maximizing profit) coalitional operation (maximizing profit ) DSM program is calculated from (1): Generation and Equilibrium between upper and consumption resources D,e D,e 0D,e D,e lower level ˜ P = P + P : (2) optimum offer itw itw itt’w itw DERs ( maximizing profit ) Retailers H-MGs with lower priority in independent or X (maximizing profit) coalitional operation (maximizing profit) 0D,e D,e D,e MCP MCP 0D,e ˜ ˜ ˜ P 6 P :  8( >  ),8P > 0 (3) itt’w itw itw tw t’w tt’w t!t 0D,e D,e Figure 2: BL-HIA structure. P > 0 :  (4) itt’w itt’w 0D,e D,e 0 Prior to the planning Planning horizon P = 0 :  ,8t = t (5) itt’w itt’w horizon Equation (3) states that if the value of the demand shifted 0 MCP MCP ˜ ˜ Selection of the upstream grid from time interval t to t is given when  >  , then tw t’w or H-MGs for coalition this value is not to exceed the predicted load value at t (i.e., Pool trading decisions in each formation and the selling D,e P ). Equations (4) and (5) are set to ensure that load shift- period of the planning horizon prices offered to the itw (H-MGs, consumers and consumers (H-MGs and ing is not defined for the same interval. The load shifted from retailers) retailers) Selection of the DER for the t to t is equivalent to the negative value of the demand de- planning horizon (H-MGs and ducted from the predicted demand at t as described by consumers) 0D,e 0D,e 0D,e P = -P :  . (6) itt’w itt’w itt’w Figure 3: Decision-making process. 4.2. CHP Objective Function and Constraints CHP,h ing this horizon. These decisions are made under uncertain- ijtw CHP CHP,e CHP,e CHP,h CHP,h FU max R = P  +P  - ij ijtw ijtw ijtw ijtw CHP,h ties of a future pool market prices, consumers’s load profile, t=1 and competitor H-MGs’s supply bids. The supply bids are a (7) function of the cost of DER installed in the H-MG. The objective is to maximize the profit that can be made 1. Consumers’s choice of energy provider: through CHP systems’s participation in DSM program as When each H-MG offers a suuply bid, consumers are to in (7): CHP,e CHP,e CHP,e CHP,e choose a H-MG as an energy provider to their electri- CHP,e P 6 P 6 P :  , , (8) ij ij ijtw ijtw ijtw cal/thermal load during the scheduling horizon. These CHP,h decisions are made based on reliable information on CHP,h CHP,h CHP,h CHP,h P 6 P 6 P :  , , (9) ij ijtw ij ijtw ijtw such prices estimated under uncertainties of pool prices CHP,h and demand. For modelling purposes, several sets of ijtw CHP,e CHP,e 0CHP,e CHP,e P =   +  : (10) consumers are created by grouping consumers with sim- ijtw ij ij ijtw CHP,h ij ilar specifications responding to H-MG’s offered prices). 2. Energy exchange in a pool market by the H-MGs: Equations (8) and (9) state upper and lower limits on the After stabilizing the H-MGs’s performance (an indepen- power generated by the CHPs. Equation (10) describes the dent or a coalitional operation) and setting supply and power generated by the CHPs as a function of the system’s demand bids, each H-MG can decide in each time inter- efficiency val of the scheduling horizon on the quantity (to/from other H-MGs in the pool market) to supply the demand 4.3. WT Objective Function and Constraints of their consumers. WT WT WT max R = P   (11) ij ijtw ijtw t=1 4. Problem Formulation The profit resulting from the participation of WT in the The H-MGs’s scheduling problem is formulated in a BL- DSM program is calculated by (11): WT WT,e HIA structure. It must be noted that dual variables have WT WT,e 0 6 P 6 P :  , (12) ijtw ij tw tw been separated by comma after equality and inequality con- straints. This section will briefly present models deployed for Equation (12) (and similar constraints for determining load shifting, and those representing the interaction among the limit on DER resources) state programmed power gen- DER, H-MGs, consumers and retailers as well as the coalition eration of DER controllable and non-controllable resources. among H-MGs. Then, BL-HIA problem formulation will be Equation (12) is related to the wind turbine electrical power presented. whose maximum limit is a parameter having uncertainty. Lower-level Upper-level 4.4. ES and TES Objective Functions and Constraints 4.7. GB Objective Function and Constraints GB ijtw GB GB GB FU max R = P   -   (25) ij ijtw ijtw ES/TES ES/TES ES/TES GB max R = P   (13) ij ij ijtw ijtw t=1 t=1 The profit made by the participation of GB in the DSM pro- gram is calculated by (25). The profit made by the participation of an electri- GB GB,h GB GB,h cal/thermal energy storage (ES/TES) system in a DSM pro- 0 6 P 6 P :  , (26) ijtw ij ijtw ijtw gram is measured by the objective function in (13): Equation (26) states the allowable limits on the heat gen- ES/TES ES/TES ES/TES ES/TES ES/TES - P 6 P 6 P :  , (14) tw eration by the gas boiler. ij ijtw ij tw ES/TES 4.8. STP Objective Function and Constraints ijtw ES/TES ES/TES ES/TES SOC = SOC + : (15) ijtw ij(t-1)w ijtw ES/TES STP STP STP max R = P   (27) ij ij ijtw ijtw t=1 ES/TES ES/TES ES/TES ES/TES,SOC ES/TES,SOC SOC 6 SOC 6 SOC :  , The profit made by the participation of STP in a DSM pro- ij ij ijtw ijtw ijtw gram is calculated by Eq. (27): (16) STP STP STP,h STP,h ES/TES ES/TES,ini 0ES/TES 0 6 P 6 P :  , , (28) SOC = SOC : (17) ijtw ij ijtw ijtw ij(t-1)w ij ijtw Equation (28) states the heat generation allowable limits ES/TES ES/TES,end 00ES/TES SOC = SOC : (18) for the operation of a thermal solar panel. In common with a ij(t=24)w ij ijw wind turbine, the solar panel thermal maximum limit is also Equations (14)–(18) are the equations governing the op- considered to be an uncertainty factor. eration of ES/TES systems. The operation of ES/TES system is subject to generation limits as in (14) and state of charge 4.9. DERs’s Price Bid Constraints X,e MCP,e limits as in (15)–(18). It should be noted that (16) states the 0 6  6  (29) ijtw tw charge/discharge rate of the ES/TES system. Y,h MCP,e 0 6  6 2  (30) tw ijtw Equations (29) and (30) are related to the electrical and 4.5. EB Constraints thermal price bids governing DERs’s operation where X in- cludes CHP, ES, WT; and Y encompasses CHP, EB, EHP, TES, EB EB,h EB,h EB,e EB,e max R = P   - P   (19) ij ijtw ijtw ijtw ijtw GB, STP. However, the value of the upper and lower bounds t=1 can vary with respect to the system deployed. The profit made by the participation of EB in the DSM pro- gram is calculated by (19): 5. H-MGs’s Independent and Coalitional Operation EB,h EB EB,e EB,h P =   P :  (20) ij tw ijtw ijtw Two scenarios are implemented to simulate the perfor- EB,h mance of the proposed BL-HIA structure. These scenarios EB,h EB,h EB,h 0 6 P 6 P :  , (21) ij ijtw ijtw ijtw are described as follows: Eqs. (20) and (21) state the consumed amount of electrical Scenario 1: power and the generated heat in the electrical boiler, respec- This scenario describes independent operations of H- tively. MGs ({A},{B},{C},{RET}). A single-level algorithm is deployed to model this scenario as further clarified by 4.6. EHP Objective Function and Constraints the independent operation of H-MGs. Scenario 2: ({A,{B,C}}, {{A,B},C}, {{A,C},B}, EHP EHP,h EHP,h EHP,e EHP,e max R = P   - P   (22) {B,{A,C}}, {{B,C},A}, {C,{A,B}}): ij ijtw ijtw ijtw ijtw t=1 This scenario describes a coalition among H-MGs taking The profit made by the participation of EHP in the DSM place at a single level of the BL-HIA structure and op- program is calculated by (22): erating in an independent operation at the other level. EHP,h EHP EHP,e HP,h P = COP  P : (23) A representation of such a scenario can take the shape ijtw ijtw ijtw of ({A,BC}, {AB,C}, {AC,B}, {B,AC}, {BC,A}, {C,{AB}}). EHP,h EHP,h EHP,h EHP,h This scenario also investigates the effect of the lower- 0 6 P 6 P :  , (24) ijtw ijtw ijtw ijtw level H-MGs forming a coalition on changing the strat- Equation (23) describes the relation between the con- egy of operating the upper-level H-MG with a high prior- sumed amount of electrical and thermal power generated by ity independently. The mathematical model of this sce- the thermal pump. In addition, (24) represents limits on the nario is further clarified by the coalitional operation of thermal generation. H-MGs as shown next. 5 5.1. Scenario 1 6. Mathematical Formulation of BL-HIA Structure Equation (31) states the profit gained by selling energy by In the upper-level problem, each H-MG seeks to maximize Retailer k to all H-MGs. its amount of profit. Each upper-level problem’s objective T n XX function states the income of each H-MG with a lower-level RET e- e- e- e- max R = (P   - P   ) (31) k kitw kitw iktw iktw priority for different scenarios. These objective functions t=1i=1 which must be maximized have been defined as the sum of the product of electrical/thermal price offers and electri- Equation (32) state the profit obtained through the inde- cal/thermal powers sold to consumers of each H-MG minus pendent operation of the H-MGs ({A}, {B},{C}). It is worth the cost of operation of DERs. mentioning that if a DER does not exist in an H-MG, it is not BL-HIA structure includes upper-level problems and a set considered in the respective objective function. 0 1 ES ES WT WT of lower-level problems under each scenario w. It is noted P   + P ijtw ijtw ijtw ijtw that if a DER is considered in the upper-level, their con- B EHP,h EHP,h C STP STP P   + P B C ijtw ijtw ijtw ijtw straints from (1)–(30) are considered in the upper-level. B C EB,h EB,h TES TES B C P   + P   similarly is the case for lower-level. The upper-level prob- B ijtw ijtw ijtw ijtw C B C T W J EB,e EB,e GB GB lem includes decision making regarding the possibility of XXX B -P   + P   C ijtw ijtw ijtw ijtw A/B/C B C R = forming coalition among H-MGs and their supply bids to B CHP,e CHP,e CHP,h CHP,h C +P   + P B C ijtw ijtw ijtw ijtw t=1w=1 j=1 achieve a higher profit. On the other hand, the quantity B C e- e- e- e- B C +P   + P by DERs/consumers resources along DSM program are in- imtw imtw iktw iktw B C B C h- h- e+ e+ cluded in the lower-level problem. It must be noted that all +P   - P @ A imtw imtw imtw imtw the power exchange among H-MGs and retailers’s decisions e+ e+ h+ h+ +P   - P iktw iktw imtw imtw are to be made on the upper-level problem. In comparison, DSM +R (i = A=B=C) decision-making variables at the lower-level include all the (32) power generation by DER resources. The upper-level objec- tive function is considered after maximizing the income of 5.2. Scenario 2 retailer or H-MGs in the case of independent or in the coali- tional operations with other H-MGs (under investigated sce- Equations (33)–(34) state the profit obtained through the narios). coalitional operation of H-MGs ({A},{B},{C}) at an upper- The income of H-MGs is defined as the product of the level or a lower-level. proposed offer for selling power to H-MGs [£/kWh] by the {AB} {A} {B} max R = R @R (33) amount of power sold to them [kWh] minus the product of a i i i power purchase offer from H-MGs [kWh] by the amount of {ABC} {A} {B} {C} max R = R @R @R (34) power bought from quantity provided by other H-MGs. i i i i where @ states the coalition among different H-MGs. In ad- 6.1. The Upper-Level Problem dition, coalitional scenario of ({B,AC}) means that the first The upper-level relationship is formulated in this section. part (B) is related to the objective function defined at the The formula expressing the DER relationship given in (1)– upper-level and a second part (AC) is related to coalition be- (30) are applicable if the related DER is considered to be in tween H-MGs A and C but defined at a lower level. the upper level. Equations (35)–(40) state the profit obtained from the coalitional operation of H-MGs A, B, and C at an upper or 6.1.1. Objective function a lower level. As was stated earlier, the objective functions of the upper- The right-hand side of these relations is made up of two level and the lower-level problems can be in the form of (32)– parts. The first part is related to the objective function de- (34). Here, the profit obtained from the coalition or the in- fined at the upper level; whereas, the second part is related dependent operation of H-MGs with more priority describes to the objective function defined at the lower level. the objective function of the upper-level problem. The upper- {A,BC} {A} {BC} max R = R @R (35) i i i level problem is to maximize the expected profit to be made by each H-MG in the case of an individual or a group opera- {AB,C} {AB} {C} max R = R @R (36) i i i tion as well as a retailer. {AC,B} {AC} {B} max R = R @R (37) i i i 6.1.2. Upper-level problem constraints (1) - (30) (41) {B,AC} {B} {AC} max R = R @R (38) i i i Equations (1)–(30) are for H-MGs with higher priority. It {BC,A} {BC} {A} max R = R @R (39) is very obvious if an H-MG with a higher priority does not i i i include any of the mentioned DERs, then the corresponding {C,AB} {C} {AB} max R = R @R (40) constraints of such DERs are to be excluded from the problem i i i 6 Heat Heat Heat Community system boundary formulation. Each H-MG in the upper-level problem makes Direct thermal connection Community’s electricity grid strategic decisions as per the following: H-MG #C EDS: Electrical distribution system EB TES CEMS FDS: Fuel distribution system Decisions of DERs on supply bid in the lower level of the Exported EMS C problem; electricity EDS H-MG #A H-MG #B Delivered Price strategic offering decisions of consumers in a H- electricity ES WT STP EHP MG on price offers. CHP CHP GB GB Delivered gas EMS A EMS B Gas Gas 6.2. The Lower-Level Problem FDS Each lower-level problem states maximizing the profit of Figure 4: The grid under study. an individual H-MG or a group of H-MGs with a lower prior- ity over different scenarios. The objective of the lower-level 6.5. Thermal Balance Constraints problem is to increase the profit of DERs. Thus, CEMS is to re- D,h CHP,h EB,h EHP,h TES GB STP duce the operating cost given limitations ruling over each of P - P + P + P + P + P + P itw ijtw ijtw ijtw ijtw ijtw ijtw the players (i.e., H-MGs, retailer, consumers). Players in BL- n n X X (45) h- h+ h HIA structure declare the amount of their generated power + P - P = 0 : tw imtw imtw and supply bids offered to the CEMS. After simulating the m=1 m=1 bilevel problem, the electrical and thermal energy prices, the Equation (45) states the relation between the thermal gen- amount of the quantity by each of the players are provided. erated and consumed power. The thermal power price in the dual variable grid corresponds to (45). After the determination of price offers related to electricity 6.2.1. Objective function and heat and also the amount of electrical and thermal power Considering the cases described in (32)–(34), the objective generation and consumption of each player, the profit made function of the lower-level problem can be taken from the by each of the players is determined. numerator of the lower-level objective function. Since each one of these lower-level problems is continu- ous and convex, it may be shown by its specific constraints 6.2.2. Lower-level problem constraints load shifting and the including Karush-Kuhn-Tucker (KKT) conditions [21]. By us- DERs ing KKT conditions, the constraints for an independent or a Equations (1)–(30) apply for each H-MG with a lower pri- coalition operation of H-MGs include the following cases: ority. It is very obvious that if an H-MG with a lower priority for example has no CHP, then its constraints must not be con-  Primal constraints (1)–(30); sidered. Equality constraints obtained from the derivative of a Lagrange expression relative to lower-level variables 6.3. Power Exchange Constraints between H-MGs and Retailers Complementary constraints obtained based on lower- Grid,e e- e- Grid,e level inequalities (3),(4)–(8)–(9)–(12)–(14)–(16)– 0 6 P 6   P :  , (42) iktw itw iktw iktw (21)–(24)- (26)–(28)–(29)–(30)–(42)–(43) e+ e+ Grid,e Grid,e 0 6 P 6   P :  , (43) iktw itw iktw iktw 7. Results and Discussion Equations (42) and (43) show the allowable limits on power exchange between retailers and H-MGs. The grid under study is shown in Figure. 4. The energy storage systems installed in H-MGs (A and C) are for storing excess electrical and thermal energy generation. The capac- 6.4. Electrical Balance Constraint ity and the number of installed equipment in each H-MG are D,e CHP,e EB,e EHP,e e- WT ES e+ shown in Table 1. P + P - P + P + P - P - P + P ijtw iktw itw ijtw ijtw ijtw ijtw iktw In Figure. 5, load profiles of H-MGs (A, B and C) are shown. n n X X e+ e+ e For an independent operation of H-MGs, most of RLDs of H- + P - P = 0 : tw imtw mitw MG A are shifted from the time intervals with higher MCP m=1 m=1 (44) to the time intervals with lower MCP. The amount of load Equation (44) states the equilibrium relation between H- shifted forms a high share of the total load of H-MGs. More MGs generated and consumed electrical power and electrical specifically, 55% of the load is shifted from time intervals with power exchange (KW) with retailers. The MCP in the grid is higher MCP to other time intervals with lower MCP taking equal to the dual variable of (44). profit maximization for H-MGs owners. On the other hand, Elec. Elec. Elec. Table 1: The Capacity and the Number of Equipment Installed in Each H-MG DERs H-MG A H-MG B H-MG C CHP Electrical output (kW ) 142 207 - ele Thermal output (kW ) 104 140 - th EHP (kW ) - 700 - th WT (kW ) 50 - - ele STP (kW ) - 600 - ele ES (kW ) (2kWh) - 500 - ele TES (m ) - - 4 GB (kW ) 2150 2150 - th EB (kW ) - - 2100 th (a) H-MG A the energy consumption in such a figure has reduced signifi- cantly when H-MGs operate in coalitional structures. Such an energy consumption is the lowest (21%) for a coalitional sce- nario of ({B,AC}). The energy consumption is at the lowest level (21%) when the coalition scenario corresponds to ({B, AC}). In addition, the reduction in the amount of load shift is a result of a DSM program aimed at achieving a higher pay- off for consumers by considering (a) employing load shifting when the value of MCP is high, along with the maximum use of H-MG A interval resources, and (b) reducing the gener- ation cost in an effective manner when load shifting is at a minimum. Moreover, the load profile of H-MGs in a coali- tion structure ({A,BC}) is the same as that of the alternative coalition structure ({AB,C}) and does not have a significant effect on the consumption level in H-MG A. This trend is com- (b) H-MG B pletely different from the case in H-MG B. More specifically, during independent operation of H-MGs, the amount of load shift in H-MG B is at its least amount (almost 30% of the total load during 24 hours). Therefore, forming a coalition among H-MGs would increase consumers’s participation in the DR program that can reach almost 42% to 50%. Such a reduction in the amount of load shift is a result of a DSM program for reaching more pay-off for consumers by considering criteria such as load shifting when the value of MCP is high, the maximum use of H-MG A interval resources, and also the reduction in the value of generation cost in the best way and with the least amount of load shifting has taken place. Alternatively, the load profile of H-MGs in a coalitional structure ({A,BC}) is the same as that of such H-MGs in an alternative coalitional structure ({AC,B}), and does not have a significant effect on the consumption nature in H-MG A. (c) H-MG C The least amount of load shifting is achieved when H-MGs Figure 5: DR+ and DR- in the H-MGs under different scenarios. B and A from a coalition in the lower-level of the BL-HIA structure, while having the objective function at an upper- level of the structure targeting maximizing the profit of H- MG C. Furthermore, these conditions are comparable for the in H-MGs independent operating conditions, is significantly {AC, B} coalition structure, having a similar nature. Under greater than the value corresponding to positive demand re- the previous conditions, a substantial share of the excess gen- sponse (DR+) conditions. While only 17% of time intervals, eration capacity is devoted to meeting H-MG C demand. As H-MG C had experienced a DR+ algorithm, such a figure a result, a negligible part of such energy has been allocated would reach 83% when a negative demand response (DR- for supplying responsive loads in H-MG B. It is important to ) would be experienced. Such a trend in demand response clarify that in the case of H-MG C, the value of the total DR-, (DR) is comparable to the scenario of coalition structures, 8 where the amount of the total load shift, with the value of DR+ total load during daily performance, are close to each other in terms of the value. The consumer’s participation per- centage in the H-MG C has improved significantly by forming a coalition between H-MG B and A reaching more than 40% of the times. Only in the coalitional structure {B,AC}, such a value can be minimum (21%). Furthermore, these conditions are also exactly similar for the {AC,B} coalitional structure and has a similar nature. Un- der the previous conditions, a big share of the amount of ex- cess generation is spent supplying H-MG C demand. It is important to clarify that for H-MG C, the value of the total DR- in H-MGs independent operation conditions is much more than such a value when positive demand re- sponse (DR+) has taken place. Such a trend in demand response (DR) is quite similar to Figure 6: H-MGs income under different scenarios. the scenario of a coalitional structures, where the amount of the total load shift and the value of DR+ total load dur- trical and thermal MCP, respectively. Although the average ing daily performance are close to each other in terms of the value of the electrical MCP in the case of an independent op- value. eration of H-MG C is at its minimum during the system’s daily The increasing trend of each H-MG income during an in- performance, such values can be significantly improved when dependent and a coalitional performance with other H-MGs investigated at individual time intervals (i.e., one hour after is shown in Figure. 6. As it is observed from this figure, each forming a coalition among H-MGs). In some of time inter- structure can be useful for one H-MG and meanwhile can vals, not only forming a coalition does not cause a degra- have no benefit for other H-MGs. The best structure, which dation in the electrical MCP, but also a small increase in can be useful for H-MG A, results from forming a coalition its value. Moreover, at certain intervals, its value may not among H-MG B and H-MG C excluding the participation of change significantly when a coalition exists compared to the H-MG A in this coalition. These conditions can also be use- scenario where H-MGs work independently. The electrical ful for H-MG B on the condition of forming a coalition with MCP value in {A,BC} coalition is about 54% of the times be- H-MG C is in a higher priority of operation. For H-MG C, come more than its value in {A,BC} combination. That is why, the highest income is experienced when this H-MG forms a no difference in values of the electrical MCP is observed for coalition with H-MG A at an initial stage given that H-MG coalitions {A,BC}, {AC,B}, {BC,A} and {C,AB}. Furthermore, B works independently. Under these conditions, the income by changing the structure from {A,BC} to {B,AC}, about 33% of H-MG A is close to the maximum value. For H-MG C, be- of the MCP value is reduced. cause of the lower the generated power, it is appropriate to Such analysis also applies for the thermal MCP for the in- form a coalition in all cases with other H-MGs. In all cases in vestigated structures. Finally, we can conclude from the sim- which H-MG C has formed a coalition with other H-MGs, an ulation results that forming a coalition among H-MGs exist- increasing trend in the income is observed. In comparison, ing in one grid will not only have a significant effect on pro- the income resulting from H-MG B when used independently, gramming and regulating the value of the power generated is significantly improved when compared to other configura- by the generation resources but also can affect the change tions, such as coalition formation with other H-MGs. Fur- in the demand consumption and the behavior of consumers thermore, it is possible, in some cases, for coalition forming participating in the DR program with a cheaper MCP. to have a detrimental effect on the H-MGs that form part of the coalition. It is also observed that the coalition among H-MG A and H- 8. Conclusion MG B at the initial level leads to a significant reduction in the income independently obtained by this H-MG. Moreover, it is desirable to prevent H-MG A from forming a coalition with The paper presented an optimum development combined H-MG B and negotiate with H-MG C to form the coalition. problem of the quantity in a deregulated electricity market In comparison, the income resulting from the independent environment. A methodology has been presented for investi- performance of H-MG B is also significant compared to other gating the possibility of increasing incomes of H-MGs, con- cases (e.g., coalition formation with other H-MGs) and in sumers and retailers existing in a multiple H-MGs. These some cases forming coalition is harmful for these H-MGs. Participants’s performances had been properly modelled in For H-MG C, because of the lowness of the generated the market environment. A H-MG programmer tries to in- power, it is appropriate to form a coalition in all cases with crease its income as long as it is freely negotiating energy ex- other H-MGs. change with DER resources and its consumers. It can also put In Figure. 7a and Figure. 7b show the values of the elec- the possibility of forming a coalition with other H-MGs on its 9 ticipation in generation and effect on electricity price, and also competition increase among H-MGs and retailers. In the upper-level problem, H-MGs change their capacity with the aim of maximizing their income and by predicting the behaviour of other competitors (H-MGs) resulting from the lower-level problem, and noting quantities and prices pro- posed by DERs and consumers. An Optimum pricing strategy was implemented to enable the market dynamic behavior on H-MGs decisions. Furthermore, a daily generation schedul- ing was presented. For a selected case study, an infinite num- ber of Nash equilibrium was observed for the case where no players tend to unilaterally change their pricing strategies. In these obtained equilibrium points, the total expected profit of all players does not change. Although it is distributed among them. (a) Electrical MCP Simulation results showed that by forming a coalition among H-MGs, their profit, the demand value of supplied load and the DERs generated power existing in those H- MGs may change. Furthermore, computational simulations showed the convergence of the proposed model for solving real problems and simultaneously presenting solutions for raising H-MGs and retailers income and also reducing market clearing price. The following results can be extracted from the structure of the developed model: The bilevel model hierarchical structure for modelling the strategic behavior of each H-MG in reaction to the behavioral change and decision making of other H-MGs and their supply bid. Furthermore, the proposed struc- ture can effectively encourage consumers to participate in the electricity market and affecting them using the DSM program. (b) Thermal MCP It had been shown that energy exchange among H-MGs Figure 7: The electrical and thermal MCP during the 24-hour system perfor- and retailers, in addition to increasing the profit of each mance. player, would have a significant impact on levelling the load and reducing consumers’s power consumption lack of support during the consumption peak period. agenda. H-MGs seek to estimate the value of the power gen- Appendices erated by DERs and also supply/demand bids to consumers. Applying KKT conditions to the lower-level problem Meanwhile, the possibility of forming coalition among H- As stated before, since the lower-level problem is a linear MGs with the aim of maximizing the income in an indepen- problem, KKT conditions can be applied to it. The constraints dent or a coalitional operation in a scheduling horizon is also obtained from the derivative of Lagrange expression relative investigated. In this way, the H-MGs encounter pool prices to lower-level variables include the following relations: uncertainties, and the value of electrical and thermal loads. Furthermore, if the supply bid of one H-MG is not competitive @L MCP,e MCP,e ˜ ˜ = 0.5  -  - D,e j 0D,e 0D,e tw t’w itw enough, consumers may choose another H-MG for supply- P >0 @P 0 itt’w itt w t =1 ing their demand. For investigating how forming a coalition D,e 0D,e 0D,e D,e min j 0D,e +  +  = 0 : itw P >0 itt’w it’tw itt’w among H-MGs can affect the market behaviour and H-MGs’s itt w (46) gained income, different scenarios were presented. These @L D,e scenarios were solved through a bilevel structure which can = +  = 0 (47) tw D,e itw @P be transformed into a one NLP problem. The proposed model itw did not only present solutions of higher income achievements @L CHP,e CHP,e CHP,e CHP,e e =  +  -  + -  = 0 (48) tw CHP,e ijtw ijtw ijtw ijtw of each H-MG in an independent or a coalitional operation, @P ijtw but also provides the higher income/lower cost for each of CHP,e CHP,e FU @L the retailers/consumers relative to a single level model. ijtw ij CHP,h CHP,h CHP,h h =  - + - - -tw = 0 CHP,h ijtw CHP,h ijtw CHP,h ijtw The BL-HIA structure has presented an adequate frame- @P N ijtw i ij work for modelling both H-MGs reaction for a better par- (49) 10 FU 25 0.14 @L A B C A B C GB,h MCP SBP SSP GB GB,h h P P P P P P t t t t,e t,e t,e t,h t,h t,h = ijtw - +  -  -  = 0 (58) tw GB GB ijtw ijtw @P N ijtw ij 0.12 TES,h @L TES,h ijtw TES TES,h h =  +  -  - -  = 0 (59) ijtw tw TES ijtw TES ijtw @P 0.1 ijtw P ij TES,h TES,SOC @L TES TES = -   +  , TES ijtw ij(t-1)w ij ijtw @SOC ijtw 0.08 (60) TES,SOC 0TES 00TES + j + j = 0 t=1 t=24 ijw ijw ijtw 0.06 The complementary constraints obtained from the inequal- ity constraints of lower-level problem are in the form of fol- lowing relations: 0.04 CHP,e CHP,e CHP,e 0 6 P - P ?  > 0 (61) ij ijtw ijtw 0 0.02 CHP,e CHP,e CHP,e 00:00 05:00 10:00 15:00 20:00 0 6 P - P ?  > 0 (62) Time [h] ijtw ij ijtw CHP,h CHP,h CHP,h Figure 8: The profiles of the electrical and thermal loads of the H-MGs and 0 6 P - P ?  > 0 (63) ij ijtw ijtw electricity prices for buying and selling. CHP,h CHP,h CHP,h 0 6 P - P ?  > 0 (64) ijtw ij ijtw 20.00 WT WT,e 0 6 P ? > 0 (65) ijtw tw WT WT,e WT 15.00 0 6 P - P ? > 0 (66) ij ijtw ijtw ES ES ES,e 0 6 P - P ?  > 0 (67) ijtw ij STP ijtw 10.00 WT ES ES ES,e 0 6 P - P ?  > 0 (68) ij ijtw ijtw ES ES ES,SOC 5.00 0 6 SOC - P ?  > 0 (69) ijtw ij ijtw ES ES ES,SOC 0 6 SOC - SOC > 0 (70) ij ijtw ijtw 0.00 EB,h EB,h 12 3 45 6 78 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 0 6 P ?  > 0 (71) ijtw ijtw Figure 9: Shape of output power wave generated by WT and STP. EB,h EB,h EB,h 0 6 P - P ?  > 0 (72) ij ijtw ijtw EHP,h EHP,h @L 0 6 P ?  > 0 (73) WT WT,e WT,e e ijtw ijtw =  + - -  = 0 (50) ijtw tw tw WT tw @P ijtw EHP,h EHP,h EHP,h 0 6 P - P ?  > 0 (74) ij ijtw ijtw ES,e @L ijtw ES ES,e ES,e e TES TES TES,h =  +  -  - -  = 0 (51) 0 6 P - P ?  > 0 (75) ijtw tw tw ES ES ijtw ij tw ijtw @P ijtw P ij TES TES,h TES 0 6 P - P ?  > 0 (76) ij ijtw ijtw @L ES,e ES,e ES,SOC ES,SOC 0ES 00ES = - + - + j + j = 0 t=1 t=24 ijw ijw ES ijtw ij(t-1)w ijtw ijtw TES TES TES,SOC @SOC ijtw 0 6 SOC - P ?  > 0 (77) ijtw ij ijtw (52) TES @L TES TES,SOC EB,h EB,h EB,h h 0 6 SOC - SOC ?  > 0 (78) + +  -  -  = 0 (53) ij ijtw ijtw tw tw EB,h ijtw ijtw @P ijtw GB GB,h 0 6 P ?  > 0 (79) ijtw ijtw @L EB,e EB EB,h e = - -   +  = 0 (54) GB ij tw tw EB,e ijtw GB GB,h @P 0 6 P - P ?  > 0 (80) ij ijtw ijtw ijtw STP STP,h @L 0 6 P ?  > 0 (81) EHP,h EHP,h EHP,h EHP,h h ijtw ijtw =  + +  -  -  = 0 (55) ijtw ijtw ijtw tw EHP,h ijtw @P ijtw STP STP,h STP 0 6 P - P ?  > 0 (82) ij ijtw ijtw @L EHP,e EHP,h EHP e = - - COP  +  = 0 (56) e+ Grid,e tw EHP,e ijtw ijtw 0 6 P ?  > 0 (83) iktw @P iktw ijtw CHP,e Grid,e WT ES e+ 0 6  P + P + P - P ?  > 0 @L ijtw ijtw iktw STP,h ijtw iktw STP STP,h h =  +  -  -  = 0 (57) ijtw tw STP ijtw ijtw (84) @P ijtw Electrical and thermal power [kW] Electricity price [£/kWh] References References [1] Kohn W, Zabinsky ZB, Nerode A. A micro-grid distributed intel- ligent control and management system. IEEE Trans Smart Grid 2015;6(6):2964–74. [2] Liu Y, Yuen C, Hassan NU, Huang S, Yu R, Xie S. Electricity cost minimization for a microgrid with distributed energy resource under different information availability. IEEE Trans Ind Electron 2015;62(4):2571–83. [3] Li C, Yu X, Yu W, Chen G, Wang J. Efficient computation for sparse load shifting in demand side management. IEEE Trans Smart Grid 2017;8(1):250–61. [4] Zazo J, Zazo S, Macua SV. Robust worst-case analysis of demand-side management in smart grids. IEEE Trans Smart Grid 2017;8(2):662– [5] Ma J, Deng J, Song L, Han Z. Incentive mechanism for demand side management in smart grid using auction. IEEE Trans Smart Grid 2014;5(3):1379–88. [6] Pei W, Deng W, Shen Z, Qi Z. Operation of battery energy storage system using extensional information model based on IEC 61850 for micro-grids. IET Gener Transm Distrib 2016;10(4):849–61. [7] Zhang Y, Zhou J, Zheng Y, Xu Y. Control optimisation for pumped stor- age unit in micro-grid with wind power penetration using improved grey wolf optimiser. IET Gener Transm Distrib 2017;11(13):3246–56. [8] Liu G, Starke M, Xiao B, Tomsovic K. Robust optimisation-based micro- grid scheduling with islanding constraints. IET Gener Transm Distrib 2017;11(7):1820–28. [9] Liu C, Wang X, Wu X, Guo J. Economic scheduling model of micro- grid considering the lifetime of batteries. IET Gener Transm Distrib 2017;11(3):759–67. [10] Panwar LK, Konda SR, Verma A, Panigrahi BK, Kumar R. Operation window constrained strategic energy management of microgrid with electric vehicle and distributed resources. IET Gener Transm Distrib 2017;11(3):615–26. [11] Kinhekar N, Padhy NP, Li F, Gupta HO. Utility oriented demand side management using smart AC and micro DC grid cooperative. IEEE Trans Power Syst 2016;31(2):1151–60. [12] Basu AK, Bhattacharya A, Chowdhury S, Chowdhury SP. Planned scheduling for economic power sharing in a CHP-based micro-grid. IEEE Trans Power Syst 2012;27(1):30–38. [13] Lee S, Kwon B, Lee S. Joint energy management system of electric supply and demand in houses and buildings. IEEE Trans Power Syst 2014;29(6):2804–12. [14] Jayaweera D. Security enhancement with nodal criticality-based integration of strategic micro grids. IEEE Trans Power Syst 2015;30(1):337–45. [15] Wang P, Gao Z, Bertling L. Operational adequacy studies of power systems with wind farms and energy storages. IEEE Trans Power Syst 2012;27(4):2377–84. [16] Giusti A, Salani M, Caro GAD, Rizzoli AE, Gambardella LM. Restricted neighborhood communication improves decentralized demand-side load management. IEEE Trans Smart Grid 2014;5(1):92–101. [17] Ni J, Ai Q. Economic power transaction using coalitional game strategy in micro-grids. IET Gener Transm Distrib 2016;10(1):10–18. [18] Gabbar HA, Othman AM. Performance optimisation for novel green plug-energy economizer in micro-grids based on recent heuristic algo- rithm. IET Gener Transm Distrib 2016;10(3):678–87. [19] Marzband M, Parhizi N, Adabi J. Optimal energy management for stand-alone microgrids based on multi-period imperialist competition algorithm considering uncertainties: experimental validation. Int Trans Electr Energy Syst 2015;30(1):122–31. [20] Valinejad J, Marzband M, Akorede MF, Barforoshi T, Jovanovic M. Gen- eration expansion planning in electricity market considering uncer- tainty in load demand and presence of strategic gencos. Electric Power Systems Research 2017;152:92–104. [21] Valinejad J, Marzband M, Akorede MF, Barforoshi T, Jovanovi´ c M. Gen- eration expansion planning in electricity market considering uncer- tainty in load demand and presence of strategic GENCOs. Electr Power Syst Res 2017;152:92–104. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Mathematics arXiv (Cornell University)

Hierarchical Interactive Architecture based on Coalition Formation for Neighborhood System Decision Making

Loading next page...
 
/lp/arxiv-cornell-university/hierarchical-interactive-architecture-based-on-coalition-formation-for-WgA7E5R1pp
ISSN
2196-5625
eISSN
ARCH-3343
DOI
10.35833/MPCE.2019.000116
Publisher site
See Article on Publisher Site

Abstract

In recent years, significant efforts for improving the technical and the economic performance of smart grids have been applied with the presence of different players making decisions in these grids. This paper proposes a bilevel hierarchical structure for designing and planning distributed energy resources in home microgrids (H-MGs). Such a small-scale grid inside the market environment permits energy exchange among distributed energy resources and home microgrids through a pool market. In this paper, each home microgrid’s planner maximizes the performance of distributed energy resources while exchanging energy with other H-MGs. The problem is formulated as a high-level problem decomposed into a set of low-level market clearing problems. The global optimal performance in terms of energy cost is met for a market structure (H-MGs, consumers, and retailers) at a Nash optimum point for the formulated scheduling game considering local and general constraints on the spot market. In general, the upper-level structure is based on H-MG generation competition for maximizing their individual and/or group income in the process of forming coalition with other H-MGs. In addition, the functionality of the lower-level of the hierarchical structure is governed by a market clearing based on the price response method by all the DERs enabling H-MGs to change the spot market strategic behavior. Prices are obtained as dual variables of power balance equations. This paper will investigate a set of bilevel games, including nonlinear programming problems solved through a complementary problem method. Using the binary theory, the bilevel hierarchical structure will be replanned using a nonlinear problem. Results prove that the proposed structure can increase the players’s profit. Keywords: Smart grids, electricity generation market, coalition formation, responsive load, bidding strategy, bilevel programming. WT wind turbine Nomenclature RET retailer Acronyms A, B, C home microgrids CHP combined heat and power Indices EB electrical boiler w the number of scenario (w 2 f1, 2, . . . ,Wg) EHP electrical heat pump i=m the number of H-MG (i 2 f1, 2, . . . ,Ig) ES+/ES- energy storage (ES) during charg- j the number of DERs existing in H-MG i (j 2 ing/discharging mode f1, 2, . . . ,Jg) GB gas boiler t time (t 2 f1, 2, . . . ,T g) k+/k- buying/ selling power from/to H-MG i/ Re- D the amount of electrical load consumer tailer k DERs’s demand (i.e., EB, EHP, ES) and/or MCP market clearing price consumers existing in H-MG i DR+, DR- amount of responsive load demand (RLD) k the number of a retailer (k 2 f1, 2, . . . ,ng) that goes from/comes to other time period e/h electrical/thermal to/from t FU fuel consumed by thermal DERs (i.e., CHP, STP solar thermal panel EB) TES thermal energy storage Email address: Jabervalinejad@vt.edu Corresponding author (Jaber Valinejad) Preprint submitted to International Journal of Electrical Power & Energy Systems October 17, 2019 arXiv:1910.07039v1 [math.OC] 15 Oct 2019 Constants 1. Introduction CHP,h the thermal efficiency of CHP ij CHP,e 0CHP,e Demand-side management (DSM) topics are focused on / the electrical efficiency of CHP ij ij energy consumption control at the consumer side [1, 2]. FU Such energy control is coordinated by electric utilities, com- the offer price of natural gas ES,ini panies and enterprises without controlling distributed energy SOC SOC initial and final value related to ES j in ij resources [3, 4]. When the latter is controlled, the topic ES,end H-MG i [kW] SOC ij would be defined as an energy management [5]. TES,ini SOC SOC initial and final value related to TES j in ij demandThis objective of this paper is to first propose a base TES,end H-MG i [kW] SOC framework for the demand of consumers encompassing H- ij EB TES ,  the thermal efficiency of EB/TES MGs, and second investigate profits that can be made from ij ij EHP COP coefficient of performance of EHP operating H-MGs independently or in a coalitional structure MC marginal cost of DER j in H-MG i [£/kWh] in a daily electricity market [6, 7]. For these purposes, en- ij ergy exchange among H-MGs is formulated as a scheduling P , P maximum electrical/thermal power gener- ij ij game. In this paper, competitive monopolies are modeled ated by DER j in H-MG i [kW] and simulated in a formulated convex optimization prob- Parameters MCP lem [8, 9]. These monopolies are based on three contra- the value of predicted electrical clearing tw dictory objectives: H-MGs’s and retailers’s income maximiza- price at time t under scenario w [£/kWh] D,e tion, consumers’s cost reduction, and demand peak reduction P predicted load power consumed by the con- itw [10]. sumers in H-MG i at time t under scenario w With the aim of a convex optimization problem, a bilevel [kW] hierarchical interactive architecture (BL-HIA) algorithm on Functions the condition of reaching a maximum profit is proposed for R the revenue of DER j related to H-MG i [£] ij both consumer and the power generator side [11, 12]. The R the revenue of H-MG i in independent oper- optimum performance problem is presented for all DERs ex- ation [£] isting in multiple H-MGs as a BL-HIA so that it is a mixed inte- im R the revenue resulting from coalitional oper- ger programming problem [13, 14]. The upper-level targets ation between H-MG i and H-MG m [£] maximizing H-MGs’s profit through energy exchange among Decision Variables H-MGs as well as H-MGs and retailers for a central optimal X,e Y,h / electrical/ thermal selling price bid by DER j ijtw ijtw performance of a decision maker [14, 15]. On the other under scenario w at time t [£/kWh] hand, the lower-level of the hierarchal structure of the prob- e h P /P electrical/thermal power generated by DER ijtw ijtw lem represents an equilibrium problem incorporating DSM j at time t under scenario w [kW] for an optimal performance of multiple H-MGs [1, 5]. This P electrical power sold to/bought from H-MG iktw way, a central optimum performance decision maker for en- i/ Retailer k at time t under scenario w [kW] ergy optimum exchange among H-MGs with each other and D,e P active power consumed by the consumers at itw with retailers in an independent and a coalitional perfor- H-MG i at time t under scenario w [kW] mance and with the aim of reaching H-MGs’s maximum profit 0D,e P shifted load power from time interval t to itt’w is included at an upper-level decision maker while consider- time interval t at H-MG i under scenario w ing an independent or a coalitional performance of all H-MGs [kW] [16–18]. The interaction between the two levels of the hier- EB,e P , consumed load power by EB/EHP j at H-MG ijtw archal structure of the game is a factor of searching for the op- EHP,e i at time t under scenario w [kW] timal solution at both levels [4, 5]. Considering the optimal ijtw MCP,e , electrical and thermal market clearing price scheduling of all H-MGs and DERs existing in them in Multi- MCP,h at time t [£/kWh] ple H-MGs requires solving mathematical program with equi- e- e+ P , P buying/selling electrical power from H-MG i librium constraints (MPEC) equivalent to bilevel problem. imtw imtw This bilevel problem can be looked at as a multiple-leader- to H-MG m [kW] e-=+ e-=+ common-follower game. The aim of implementing this game P /P buying/ selling electrical power from/to H- iktw kitw structure is finding a final equilibrium point in which none MG i/ Retailer k [kW] h- h+ of H-MGs and consumers can increase their profit by chang- P /P buying/selling thermal power from/to H- imtw imtw ing in the generation and the consumption schedules. Fur- MG i/H-MG m [kW] e- e+ thermore, the BL-HIA structure accounts for decisions result- / supply bids for buying/ selling electrical iktw iktw ing from forming a coalition among H-MGs to maximize the power from/to H-MG i/Retailer k [kW] h- h+ profit and also exchange energy among them. / supply bids for buying/selling thermal power imtw imtw The contribution in this work can be summarized as fol- from/to H-MG i/H-MG m [kW] lows: First, the proposed BL-HIA structure is preferable over the proposed structure in [19] as it is a multi-ownership structure that permits, forming coalition among H-MGs and D E R D ER D ER DER DER DER DER FL:A...SL:BC H- H- FL:B...SL:AC MG MG FL:C...SL:AB DE R FL:AB...SL:C FL:AC...SL:B FL:BC...SL:A D E R H- MG CEMS EMS DER:CHP-WT-STP-EB-EHP-ES- GB UL: upper level LL: lower leveL H-MG A and B :AB D ER DER DER H- H- MG MG D E R H- MG CEMS EMS DER:CHP-WT-STP-EB-EHP-ES- GB UL: upper level LL: lower leveL H-MG A and B :AB FL:A...SL:BC FL:B...SL:AC FL:C...SL:AB FL:AB...SL:C FL:AC...SL:B FL:BC...SL:A structure that is a decision-making problem including sev- variables DERs ,consumers , DER:CHP-WT-STP- load shifting of building with EB-EHP-ES-GB eral agents which try to optimize their corresponding objec- UL: upper level higher priority LL: lower level tive functions on a connectable dependent set. An agent is, AB:Coalition between DER H-MG A and B in fact, an object which can act as a DER or connected to DER profit profit other units. The BL-HIA structure is shown in Figure. 2. The maximization of the maximization of the upper-level problem states maximizing the profit of H-MGs profit of each H-MG profit of each H-MG Coalition with higher priority with higher priority having a higher priority of operating in an independent or a UL:A...LL:BC load shifting load shifting coalitional operation on the condition of: 1) satisfying upper- UL:B...LL:AC profit profit UL:C...LL:AB level constraints and 2) satisfying a set of lower-level prob- Upper- Lower- UL:AB...LL:C DERs variables ,consumers ,the H-MG ,load lems. H-MGs with a higher priority of operation in the upper- shifting with lower priority ,electricity prices UL:AC...LL:B level level and market hourly heat UL:BC...LL:A CEMS EMS level problem are identified based on their price bids. Upper- level constraints include limits on the quantity and supply bids of DERs resources, the minimum accessible power ca- Figure 1: The proposed BL-HIA structure illustrating a variety of coali- pacity by the market regulator, and buying/selling quantities tion formation among H-MGs. UL:A...LL:BC implies that the HMG A is modeled in the upper level while HMGs B and C are modeled in lower by H-MGs and retailers. Lower-level problem states the mar- level.Distributed energy resources encompass set of distributed generation, ket clearing prices (MCP) with the aim of maximizing the electrical and thermal energy storage. profit of H-MGs having lower priority of operation subject to meeting equilibrium constraints for each H-MG, genera- tion/consumption limits, and the number of consumers par- explicitly increasing the competition among H-MGs and con- ticipating in the DSM program. sumers rather than an independent operation of H-MGs. Sec- As it is observed in Figure. 2, the higher priority of H- ond, the BL-HIA is adequate for modelling problems with MGs operation (in an independent or a coalitional opera- several leaders (i.e., H-MGs) having their own individual ob- tion) is defined by the energy excess/shortage gap and sup- jective functions when operating independently or in a coali- ply/demand bids that permits a maximization of an expected tional manner (upper-level problem). Such a game is to opti- profit. Maximizing the profit of each H-MG is achieved by mize several followers (i.e., consumers inserted in the bilevel considering the fact that each agent at the lower-level prob- structure). These models are related to situations where ac- lem shows an optimal operation in correspondence to the in- tions and followers’s performance in BL-HIA have a signifi- come of H-MGs with higher priority of operation based on cant effect on decisions made by leaders. This fact is related the offered price. This optimal operation includes an estima- to the case in which H-MGs’s profit (a leader) depends on tion of the demand supplied and shifted by each consumer the amount of energy which is sold for supplying consumers in an independent or a coalitional operation. It must be em- existing in the power grid (as follower). The general view phasized that the competition among H-MGs with higher and of the hierarchical structure and optimization problems has lower priority (or competitor H-MGs) are explicitly modelled been shown with the proposed model in Figure. 2. Third, a at upper and lower-levels. It must be noted that upper-level better strategy maximizing consumers’s satisfaction in terms and lower-level problems shown in Figure. 2 become related of demand supply and H-MGs’s profit is presented in compar- to each other. In other words, lower-level problems estimate ison to a single level structure. Finally, the BL-HIA structure is the price and the quantity of competitor H-MGs which di- solved by formulating an equivalent one level mixed-integer rectly affect the profit of H-MGs inserted in the above prob- nonlinear programming (MINLP) problem deploying the KKT lem. In other words, decisions related to forming a coalition optimization conditions. The innovations in this paper can be and taking a bidding strategy by the competitor strategic H- summarized as follows: MGs in the upper-level problem have also a significant effect 1. An optimum programming solution within H-MGs gen- on the MCP resulted from the lower-level problem. eration as a BL-HIA structure; The BL-HIA structure is shown in Figure. 1. In BL-HIA 2. Providing a multiple-leader-common-follower game structure is also considered for modelling the uncertainty of which states the effectiveness of the market competition pool prices, electrical/thermal load demand, and H-MGs buy- in multiple H-MGs through solving a BL-HIA structure; ing/selling prices. The proposed bilevel model has been sim- 3. Developing a new model for demand-side management; plified as a one-level problem using KKT method [20] for the 4. Accommodating both DR resources and storage devices sake of convexification. in the market operation to achieve a comprehensive so- lution exploiting all flexibilities; and 3. Decision-Making Process by BL-HIA Structure 5. Proposing an advanced electricity market for active dis- tribution networks based on game theory; The decision-making process in the BL-HIA structure of H- MGs, consumers and retailers can be summarized as shown in Figure. 3. At the beginning of the scheduling horizon, each 2. The Proposed BL-HIA Structure H-MG presents the necessary decisions on DSM for an inde- The problem encountered by the H-MGs for an indepen- pendent or a coalitional operation with other H-MGs. More- dent or a coalitional operation can be modelled as a bilevel over, supply/demand bids are provided to the consumers dur- 3 Wind speed Solar irradiance 4.1. DR Objective Function and Constraints prediction prediction Power Power Electricity price generated by Load forcasting D,e DSM MCP MCP generated by PV prediction ˜ ˜ WT max R = 0.5 P   -  (1) i it’w tw t’w Modelling uncertainty using Taghochi method 0 t =1 Retailers H-MGs with higher priority in independent or The profit resulting from the participation of consumers in (maximizing profit) coalitional operation (maximizing profit ) DSM program is calculated from (1): Generation and Equilibrium between upper and consumption resources D,e D,e 0D,e D,e lower level ˜ P = P + P : (2) optimum offer itw itw itt’w itw DERs ( maximizing profit ) Retailers H-MGs with lower priority in independent or X (maximizing profit) coalitional operation (maximizing profit) 0D,e D,e D,e MCP MCP 0D,e ˜ ˜ ˜ P 6 P :  8( >  ),8P > 0 (3) itt’w itw itw tw t’w tt’w t!t 0D,e D,e Figure 2: BL-HIA structure. P > 0 :  (4) itt’w itt’w 0D,e D,e 0 Prior to the planning Planning horizon P = 0 :  ,8t = t (5) itt’w itt’w horizon Equation (3) states that if the value of the demand shifted 0 MCP MCP ˜ ˜ Selection of the upstream grid from time interval t to t is given when  >  , then tw t’w or H-MGs for coalition this value is not to exceed the predicted load value at t (i.e., Pool trading decisions in each formation and the selling D,e P ). Equations (4) and (5) are set to ensure that load shift- period of the planning horizon prices offered to the itw (H-MGs, consumers and consumers (H-MGs and ing is not defined for the same interval. The load shifted from retailers) retailers) Selection of the DER for the t to t is equivalent to the negative value of the demand de- planning horizon (H-MGs and ducted from the predicted demand at t as described by consumers) 0D,e 0D,e 0D,e P = -P :  . (6) itt’w itt’w itt’w Figure 3: Decision-making process. 4.2. CHP Objective Function and Constraints CHP,h ing this horizon. These decisions are made under uncertain- ijtw CHP CHP,e CHP,e CHP,h CHP,h FU max R = P  +P  - ij ijtw ijtw ijtw ijtw CHP,h ties of a future pool market prices, consumers’s load profile, t=1 and competitor H-MGs’s supply bids. The supply bids are a (7) function of the cost of DER installed in the H-MG. The objective is to maximize the profit that can be made 1. Consumers’s choice of energy provider: through CHP systems’s participation in DSM program as When each H-MG offers a suuply bid, consumers are to in (7): CHP,e CHP,e CHP,e CHP,e choose a H-MG as an energy provider to their electri- CHP,e P 6 P 6 P :  , , (8) ij ij ijtw ijtw ijtw cal/thermal load during the scheduling horizon. These CHP,h decisions are made based on reliable information on CHP,h CHP,h CHP,h CHP,h P 6 P 6 P :  , , (9) ij ijtw ij ijtw ijtw such prices estimated under uncertainties of pool prices CHP,h and demand. For modelling purposes, several sets of ijtw CHP,e CHP,e 0CHP,e CHP,e P =   +  : (10) consumers are created by grouping consumers with sim- ijtw ij ij ijtw CHP,h ij ilar specifications responding to H-MG’s offered prices). 2. Energy exchange in a pool market by the H-MGs: Equations (8) and (9) state upper and lower limits on the After stabilizing the H-MGs’s performance (an indepen- power generated by the CHPs. Equation (10) describes the dent or a coalitional operation) and setting supply and power generated by the CHPs as a function of the system’s demand bids, each H-MG can decide in each time inter- efficiency val of the scheduling horizon on the quantity (to/from other H-MGs in the pool market) to supply the demand 4.3. WT Objective Function and Constraints of their consumers. WT WT WT max R = P   (11) ij ijtw ijtw t=1 4. Problem Formulation The profit resulting from the participation of WT in the The H-MGs’s scheduling problem is formulated in a BL- DSM program is calculated by (11): WT WT,e HIA structure. It must be noted that dual variables have WT WT,e 0 6 P 6 P :  , (12) ijtw ij tw tw been separated by comma after equality and inequality con- straints. This section will briefly present models deployed for Equation (12) (and similar constraints for determining load shifting, and those representing the interaction among the limit on DER resources) state programmed power gen- DER, H-MGs, consumers and retailers as well as the coalition eration of DER controllable and non-controllable resources. among H-MGs. Then, BL-HIA problem formulation will be Equation (12) is related to the wind turbine electrical power presented. whose maximum limit is a parameter having uncertainty. Lower-level Upper-level 4.4. ES and TES Objective Functions and Constraints 4.7. GB Objective Function and Constraints GB ijtw GB GB GB FU max R = P   -   (25) ij ijtw ijtw ES/TES ES/TES ES/TES GB max R = P   (13) ij ij ijtw ijtw t=1 t=1 The profit made by the participation of GB in the DSM pro- gram is calculated by (25). The profit made by the participation of an electri- GB GB,h GB GB,h cal/thermal energy storage (ES/TES) system in a DSM pro- 0 6 P 6 P :  , (26) ijtw ij ijtw ijtw gram is measured by the objective function in (13): Equation (26) states the allowable limits on the heat gen- ES/TES ES/TES ES/TES ES/TES ES/TES - P 6 P 6 P :  , (14) tw eration by the gas boiler. ij ijtw ij tw ES/TES 4.8. STP Objective Function and Constraints ijtw ES/TES ES/TES ES/TES SOC = SOC + : (15) ijtw ij(t-1)w ijtw ES/TES STP STP STP max R = P   (27) ij ij ijtw ijtw t=1 ES/TES ES/TES ES/TES ES/TES,SOC ES/TES,SOC SOC 6 SOC 6 SOC :  , The profit made by the participation of STP in a DSM pro- ij ij ijtw ijtw ijtw gram is calculated by Eq. (27): (16) STP STP STP,h STP,h ES/TES ES/TES,ini 0ES/TES 0 6 P 6 P :  , , (28) SOC = SOC : (17) ijtw ij ijtw ijtw ij(t-1)w ij ijtw Equation (28) states the heat generation allowable limits ES/TES ES/TES,end 00ES/TES SOC = SOC : (18) for the operation of a thermal solar panel. In common with a ij(t=24)w ij ijw wind turbine, the solar panel thermal maximum limit is also Equations (14)–(18) are the equations governing the op- considered to be an uncertainty factor. eration of ES/TES systems. The operation of ES/TES system is subject to generation limits as in (14) and state of charge 4.9. DERs’s Price Bid Constraints X,e MCP,e limits as in (15)–(18). It should be noted that (16) states the 0 6  6  (29) ijtw tw charge/discharge rate of the ES/TES system. Y,h MCP,e 0 6  6 2  (30) tw ijtw Equations (29) and (30) are related to the electrical and 4.5. EB Constraints thermal price bids governing DERs’s operation where X in- cludes CHP, ES, WT; and Y encompasses CHP, EB, EHP, TES, EB EB,h EB,h EB,e EB,e max R = P   - P   (19) ij ijtw ijtw ijtw ijtw GB, STP. However, the value of the upper and lower bounds t=1 can vary with respect to the system deployed. The profit made by the participation of EB in the DSM pro- gram is calculated by (19): 5. H-MGs’s Independent and Coalitional Operation EB,h EB EB,e EB,h P =   P :  (20) ij tw ijtw ijtw Two scenarios are implemented to simulate the perfor- EB,h mance of the proposed BL-HIA structure. These scenarios EB,h EB,h EB,h 0 6 P 6 P :  , (21) ij ijtw ijtw ijtw are described as follows: Eqs. (20) and (21) state the consumed amount of electrical Scenario 1: power and the generated heat in the electrical boiler, respec- This scenario describes independent operations of H- tively. MGs ({A},{B},{C},{RET}). A single-level algorithm is deployed to model this scenario as further clarified by 4.6. EHP Objective Function and Constraints the independent operation of H-MGs. Scenario 2: ({A,{B,C}}, {{A,B},C}, {{A,C},B}, EHP EHP,h EHP,h EHP,e EHP,e max R = P   - P   (22) {B,{A,C}}, {{B,C},A}, {C,{A,B}}): ij ijtw ijtw ijtw ijtw t=1 This scenario describes a coalition among H-MGs taking The profit made by the participation of EHP in the DSM place at a single level of the BL-HIA structure and op- program is calculated by (22): erating in an independent operation at the other level. EHP,h EHP EHP,e HP,h P = COP  P : (23) A representation of such a scenario can take the shape ijtw ijtw ijtw of ({A,BC}, {AB,C}, {AC,B}, {B,AC}, {BC,A}, {C,{AB}}). EHP,h EHP,h EHP,h EHP,h This scenario also investigates the effect of the lower- 0 6 P 6 P :  , (24) ijtw ijtw ijtw ijtw level H-MGs forming a coalition on changing the strat- Equation (23) describes the relation between the con- egy of operating the upper-level H-MG with a high prior- sumed amount of electrical and thermal power generated by ity independently. The mathematical model of this sce- the thermal pump. In addition, (24) represents limits on the nario is further clarified by the coalitional operation of thermal generation. H-MGs as shown next. 5 5.1. Scenario 1 6. Mathematical Formulation of BL-HIA Structure Equation (31) states the profit gained by selling energy by In the upper-level problem, each H-MG seeks to maximize Retailer k to all H-MGs. its amount of profit. Each upper-level problem’s objective T n XX function states the income of each H-MG with a lower-level RET e- e- e- e- max R = (P   - P   ) (31) k kitw kitw iktw iktw priority for different scenarios. These objective functions t=1i=1 which must be maximized have been defined as the sum of the product of electrical/thermal price offers and electri- Equation (32) state the profit obtained through the inde- cal/thermal powers sold to consumers of each H-MG minus pendent operation of the H-MGs ({A}, {B},{C}). It is worth the cost of operation of DERs. mentioning that if a DER does not exist in an H-MG, it is not BL-HIA structure includes upper-level problems and a set considered in the respective objective function. 0 1 ES ES WT WT of lower-level problems under each scenario w. It is noted P   + P ijtw ijtw ijtw ijtw that if a DER is considered in the upper-level, their con- B EHP,h EHP,h C STP STP P   + P B C ijtw ijtw ijtw ijtw straints from (1)–(30) are considered in the upper-level. B C EB,h EB,h TES TES B C P   + P   similarly is the case for lower-level. The upper-level prob- B ijtw ijtw ijtw ijtw C B C T W J EB,e EB,e GB GB lem includes decision making regarding the possibility of XXX B -P   + P   C ijtw ijtw ijtw ijtw A/B/C B C R = forming coalition among H-MGs and their supply bids to B CHP,e CHP,e CHP,h CHP,h C +P   + P B C ijtw ijtw ijtw ijtw t=1w=1 j=1 achieve a higher profit. On the other hand, the quantity B C e- e- e- e- B C +P   + P by DERs/consumers resources along DSM program are in- imtw imtw iktw iktw B C B C h- h- e+ e+ cluded in the lower-level problem. It must be noted that all +P   - P @ A imtw imtw imtw imtw the power exchange among H-MGs and retailers’s decisions e+ e+ h+ h+ +P   - P iktw iktw imtw imtw are to be made on the upper-level problem. In comparison, DSM +R (i = A=B=C) decision-making variables at the lower-level include all the (32) power generation by DER resources. The upper-level objec- tive function is considered after maximizing the income of 5.2. Scenario 2 retailer or H-MGs in the case of independent or in the coali- tional operations with other H-MGs (under investigated sce- Equations (33)–(34) state the profit obtained through the narios). coalitional operation of H-MGs ({A},{B},{C}) at an upper- The income of H-MGs is defined as the product of the level or a lower-level. proposed offer for selling power to H-MGs [£/kWh] by the {AB} {A} {B} max R = R @R (33) amount of power sold to them [kWh] minus the product of a i i i power purchase offer from H-MGs [kWh] by the amount of {ABC} {A} {B} {C} max R = R @R @R (34) power bought from quantity provided by other H-MGs. i i i i where @ states the coalition among different H-MGs. In ad- 6.1. The Upper-Level Problem dition, coalitional scenario of ({B,AC}) means that the first The upper-level relationship is formulated in this section. part (B) is related to the objective function defined at the The formula expressing the DER relationship given in (1)– upper-level and a second part (AC) is related to coalition be- (30) are applicable if the related DER is considered to be in tween H-MGs A and C but defined at a lower level. the upper level. Equations (35)–(40) state the profit obtained from the coalitional operation of H-MGs A, B, and C at an upper or 6.1.1. Objective function a lower level. As was stated earlier, the objective functions of the upper- The right-hand side of these relations is made up of two level and the lower-level problems can be in the form of (32)– parts. The first part is related to the objective function de- (34). Here, the profit obtained from the coalition or the in- fined at the upper level; whereas, the second part is related dependent operation of H-MGs with more priority describes to the objective function defined at the lower level. the objective function of the upper-level problem. The upper- {A,BC} {A} {BC} max R = R @R (35) i i i level problem is to maximize the expected profit to be made by each H-MG in the case of an individual or a group opera- {AB,C} {AB} {C} max R = R @R (36) i i i tion as well as a retailer. {AC,B} {AC} {B} max R = R @R (37) i i i 6.1.2. Upper-level problem constraints (1) - (30) (41) {B,AC} {B} {AC} max R = R @R (38) i i i Equations (1)–(30) are for H-MGs with higher priority. It {BC,A} {BC} {A} max R = R @R (39) is very obvious if an H-MG with a higher priority does not i i i include any of the mentioned DERs, then the corresponding {C,AB} {C} {AB} max R = R @R (40) constraints of such DERs are to be excluded from the problem i i i 6 Heat Heat Heat Community system boundary formulation. Each H-MG in the upper-level problem makes Direct thermal connection Community’s electricity grid strategic decisions as per the following: H-MG #C EDS: Electrical distribution system EB TES CEMS FDS: Fuel distribution system Decisions of DERs on supply bid in the lower level of the Exported EMS C problem; electricity EDS H-MG #A H-MG #B Delivered Price strategic offering decisions of consumers in a H- electricity ES WT STP EHP MG on price offers. CHP CHP GB GB Delivered gas EMS A EMS B Gas Gas 6.2. The Lower-Level Problem FDS Each lower-level problem states maximizing the profit of Figure 4: The grid under study. an individual H-MG or a group of H-MGs with a lower prior- ity over different scenarios. The objective of the lower-level 6.5. Thermal Balance Constraints problem is to increase the profit of DERs. Thus, CEMS is to re- D,h CHP,h EB,h EHP,h TES GB STP duce the operating cost given limitations ruling over each of P - P + P + P + P + P + P itw ijtw ijtw ijtw ijtw ijtw ijtw the players (i.e., H-MGs, retailer, consumers). Players in BL- n n X X (45) h- h+ h HIA structure declare the amount of their generated power + P - P = 0 : tw imtw imtw and supply bids offered to the CEMS. After simulating the m=1 m=1 bilevel problem, the electrical and thermal energy prices, the Equation (45) states the relation between the thermal gen- amount of the quantity by each of the players are provided. erated and consumed power. The thermal power price in the dual variable grid corresponds to (45). After the determination of price offers related to electricity 6.2.1. Objective function and heat and also the amount of electrical and thermal power Considering the cases described in (32)–(34), the objective generation and consumption of each player, the profit made function of the lower-level problem can be taken from the by each of the players is determined. numerator of the lower-level objective function. Since each one of these lower-level problems is continu- ous and convex, it may be shown by its specific constraints 6.2.2. Lower-level problem constraints load shifting and the including Karush-Kuhn-Tucker (KKT) conditions [21]. By us- DERs ing KKT conditions, the constraints for an independent or a Equations (1)–(30) apply for each H-MG with a lower pri- coalition operation of H-MGs include the following cases: ority. It is very obvious that if an H-MG with a lower priority for example has no CHP, then its constraints must not be con-  Primal constraints (1)–(30); sidered. Equality constraints obtained from the derivative of a Lagrange expression relative to lower-level variables 6.3. Power Exchange Constraints between H-MGs and Retailers Complementary constraints obtained based on lower- Grid,e e- e- Grid,e level inequalities (3),(4)–(8)–(9)–(12)–(14)–(16)– 0 6 P 6   P :  , (42) iktw itw iktw iktw (21)–(24)- (26)–(28)–(29)–(30)–(42)–(43) e+ e+ Grid,e Grid,e 0 6 P 6   P :  , (43) iktw itw iktw iktw 7. Results and Discussion Equations (42) and (43) show the allowable limits on power exchange between retailers and H-MGs. The grid under study is shown in Figure. 4. The energy storage systems installed in H-MGs (A and C) are for storing excess electrical and thermal energy generation. The capac- 6.4. Electrical Balance Constraint ity and the number of installed equipment in each H-MG are D,e CHP,e EB,e EHP,e e- WT ES e+ shown in Table 1. P + P - P + P + P - P - P + P ijtw iktw itw ijtw ijtw ijtw ijtw iktw In Figure. 5, load profiles of H-MGs (A, B and C) are shown. n n X X e+ e+ e For an independent operation of H-MGs, most of RLDs of H- + P - P = 0 : tw imtw mitw MG A are shifted from the time intervals with higher MCP m=1 m=1 (44) to the time intervals with lower MCP. The amount of load Equation (44) states the equilibrium relation between H- shifted forms a high share of the total load of H-MGs. More MGs generated and consumed electrical power and electrical specifically, 55% of the load is shifted from time intervals with power exchange (KW) with retailers. The MCP in the grid is higher MCP to other time intervals with lower MCP taking equal to the dual variable of (44). profit maximization for H-MGs owners. On the other hand, Elec. Elec. Elec. Table 1: The Capacity and the Number of Equipment Installed in Each H-MG DERs H-MG A H-MG B H-MG C CHP Electrical output (kW ) 142 207 - ele Thermal output (kW ) 104 140 - th EHP (kW ) - 700 - th WT (kW ) 50 - - ele STP (kW ) - 600 - ele ES (kW ) (2kWh) - 500 - ele TES (m ) - - 4 GB (kW ) 2150 2150 - th EB (kW ) - - 2100 th (a) H-MG A the energy consumption in such a figure has reduced signifi- cantly when H-MGs operate in coalitional structures. Such an energy consumption is the lowest (21%) for a coalitional sce- nario of ({B,AC}). The energy consumption is at the lowest level (21%) when the coalition scenario corresponds to ({B, AC}). In addition, the reduction in the amount of load shift is a result of a DSM program aimed at achieving a higher pay- off for consumers by considering (a) employing load shifting when the value of MCP is high, along with the maximum use of H-MG A interval resources, and (b) reducing the gener- ation cost in an effective manner when load shifting is at a minimum. Moreover, the load profile of H-MGs in a coali- tion structure ({A,BC}) is the same as that of the alternative coalition structure ({AB,C}) and does not have a significant effect on the consumption level in H-MG A. This trend is com- (b) H-MG B pletely different from the case in H-MG B. More specifically, during independent operation of H-MGs, the amount of load shift in H-MG B is at its least amount (almost 30% of the total load during 24 hours). Therefore, forming a coalition among H-MGs would increase consumers’s participation in the DR program that can reach almost 42% to 50%. Such a reduction in the amount of load shift is a result of a DSM program for reaching more pay-off for consumers by considering criteria such as load shifting when the value of MCP is high, the maximum use of H-MG A interval resources, and also the reduction in the value of generation cost in the best way and with the least amount of load shifting has taken place. Alternatively, the load profile of H-MGs in a coalitional structure ({A,BC}) is the same as that of such H-MGs in an alternative coalitional structure ({AC,B}), and does not have a significant effect on the consumption nature in H-MG A. (c) H-MG C The least amount of load shifting is achieved when H-MGs Figure 5: DR+ and DR- in the H-MGs under different scenarios. B and A from a coalition in the lower-level of the BL-HIA structure, while having the objective function at an upper- level of the structure targeting maximizing the profit of H- MG C. Furthermore, these conditions are comparable for the in H-MGs independent operating conditions, is significantly {AC, B} coalition structure, having a similar nature. Under greater than the value corresponding to positive demand re- the previous conditions, a substantial share of the excess gen- sponse (DR+) conditions. While only 17% of time intervals, eration capacity is devoted to meeting H-MG C demand. As H-MG C had experienced a DR+ algorithm, such a figure a result, a negligible part of such energy has been allocated would reach 83% when a negative demand response (DR- for supplying responsive loads in H-MG B. It is important to ) would be experienced. Such a trend in demand response clarify that in the case of H-MG C, the value of the total DR-, (DR) is comparable to the scenario of coalition structures, 8 where the amount of the total load shift, with the value of DR+ total load during daily performance, are close to each other in terms of the value. The consumer’s participation per- centage in the H-MG C has improved significantly by forming a coalition between H-MG B and A reaching more than 40% of the times. Only in the coalitional structure {B,AC}, such a value can be minimum (21%). Furthermore, these conditions are also exactly similar for the {AC,B} coalitional structure and has a similar nature. Un- der the previous conditions, a big share of the amount of ex- cess generation is spent supplying H-MG C demand. It is important to clarify that for H-MG C, the value of the total DR- in H-MGs independent operation conditions is much more than such a value when positive demand re- sponse (DR+) has taken place. Such a trend in demand response (DR) is quite similar to Figure 6: H-MGs income under different scenarios. the scenario of a coalitional structures, where the amount of the total load shift and the value of DR+ total load dur- trical and thermal MCP, respectively. Although the average ing daily performance are close to each other in terms of the value of the electrical MCP in the case of an independent op- value. eration of H-MG C is at its minimum during the system’s daily The increasing trend of each H-MG income during an in- performance, such values can be significantly improved when dependent and a coalitional performance with other H-MGs investigated at individual time intervals (i.e., one hour after is shown in Figure. 6. As it is observed from this figure, each forming a coalition among H-MGs). In some of time inter- structure can be useful for one H-MG and meanwhile can vals, not only forming a coalition does not cause a degra- have no benefit for other H-MGs. The best structure, which dation in the electrical MCP, but also a small increase in can be useful for H-MG A, results from forming a coalition its value. Moreover, at certain intervals, its value may not among H-MG B and H-MG C excluding the participation of change significantly when a coalition exists compared to the H-MG A in this coalition. These conditions can also be use- scenario where H-MGs work independently. The electrical ful for H-MG B on the condition of forming a coalition with MCP value in {A,BC} coalition is about 54% of the times be- H-MG C is in a higher priority of operation. For H-MG C, come more than its value in {A,BC} combination. That is why, the highest income is experienced when this H-MG forms a no difference in values of the electrical MCP is observed for coalition with H-MG A at an initial stage given that H-MG coalitions {A,BC}, {AC,B}, {BC,A} and {C,AB}. Furthermore, B works independently. Under these conditions, the income by changing the structure from {A,BC} to {B,AC}, about 33% of H-MG A is close to the maximum value. For H-MG C, be- of the MCP value is reduced. cause of the lower the generated power, it is appropriate to Such analysis also applies for the thermal MCP for the in- form a coalition in all cases with other H-MGs. In all cases in vestigated structures. Finally, we can conclude from the sim- which H-MG C has formed a coalition with other H-MGs, an ulation results that forming a coalition among H-MGs exist- increasing trend in the income is observed. In comparison, ing in one grid will not only have a significant effect on pro- the income resulting from H-MG B when used independently, gramming and regulating the value of the power generated is significantly improved when compared to other configura- by the generation resources but also can affect the change tions, such as coalition formation with other H-MGs. Fur- in the demand consumption and the behavior of consumers thermore, it is possible, in some cases, for coalition forming participating in the DR program with a cheaper MCP. to have a detrimental effect on the H-MGs that form part of the coalition. It is also observed that the coalition among H-MG A and H- 8. Conclusion MG B at the initial level leads to a significant reduction in the income independently obtained by this H-MG. Moreover, it is desirable to prevent H-MG A from forming a coalition with The paper presented an optimum development combined H-MG B and negotiate with H-MG C to form the coalition. problem of the quantity in a deregulated electricity market In comparison, the income resulting from the independent environment. A methodology has been presented for investi- performance of H-MG B is also significant compared to other gating the possibility of increasing incomes of H-MGs, con- cases (e.g., coalition formation with other H-MGs) and in sumers and retailers existing in a multiple H-MGs. These some cases forming coalition is harmful for these H-MGs. Participants’s performances had been properly modelled in For H-MG C, because of the lowness of the generated the market environment. A H-MG programmer tries to in- power, it is appropriate to form a coalition in all cases with crease its income as long as it is freely negotiating energy ex- other H-MGs. change with DER resources and its consumers. It can also put In Figure. 7a and Figure. 7b show the values of the elec- the possibility of forming a coalition with other H-MGs on its 9 ticipation in generation and effect on electricity price, and also competition increase among H-MGs and retailers. In the upper-level problem, H-MGs change their capacity with the aim of maximizing their income and by predicting the behaviour of other competitors (H-MGs) resulting from the lower-level problem, and noting quantities and prices pro- posed by DERs and consumers. An Optimum pricing strategy was implemented to enable the market dynamic behavior on H-MGs decisions. Furthermore, a daily generation schedul- ing was presented. For a selected case study, an infinite num- ber of Nash equilibrium was observed for the case where no players tend to unilaterally change their pricing strategies. In these obtained equilibrium points, the total expected profit of all players does not change. Although it is distributed among them. (a) Electrical MCP Simulation results showed that by forming a coalition among H-MGs, their profit, the demand value of supplied load and the DERs generated power existing in those H- MGs may change. Furthermore, computational simulations showed the convergence of the proposed model for solving real problems and simultaneously presenting solutions for raising H-MGs and retailers income and also reducing market clearing price. The following results can be extracted from the structure of the developed model: The bilevel model hierarchical structure for modelling the strategic behavior of each H-MG in reaction to the behavioral change and decision making of other H-MGs and their supply bid. Furthermore, the proposed struc- ture can effectively encourage consumers to participate in the electricity market and affecting them using the DSM program. (b) Thermal MCP It had been shown that energy exchange among H-MGs Figure 7: The electrical and thermal MCP during the 24-hour system perfor- and retailers, in addition to increasing the profit of each mance. player, would have a significant impact on levelling the load and reducing consumers’s power consumption lack of support during the consumption peak period. agenda. H-MGs seek to estimate the value of the power gen- Appendices erated by DERs and also supply/demand bids to consumers. Applying KKT conditions to the lower-level problem Meanwhile, the possibility of forming coalition among H- As stated before, since the lower-level problem is a linear MGs with the aim of maximizing the income in an indepen- problem, KKT conditions can be applied to it. The constraints dent or a coalitional operation in a scheduling horizon is also obtained from the derivative of Lagrange expression relative investigated. In this way, the H-MGs encounter pool prices to lower-level variables include the following relations: uncertainties, and the value of electrical and thermal loads. Furthermore, if the supply bid of one H-MG is not competitive @L MCP,e MCP,e ˜ ˜ = 0.5  -  - D,e j 0D,e 0D,e tw t’w itw enough, consumers may choose another H-MG for supply- P >0 @P 0 itt’w itt w t =1 ing their demand. For investigating how forming a coalition D,e 0D,e 0D,e D,e min j 0D,e +  +  = 0 : itw P >0 itt’w it’tw itt’w among H-MGs can affect the market behaviour and H-MGs’s itt w (46) gained income, different scenarios were presented. These @L D,e scenarios were solved through a bilevel structure which can = +  = 0 (47) tw D,e itw @P be transformed into a one NLP problem. The proposed model itw did not only present solutions of higher income achievements @L CHP,e CHP,e CHP,e CHP,e e =  +  -  + -  = 0 (48) tw CHP,e ijtw ijtw ijtw ijtw of each H-MG in an independent or a coalitional operation, @P ijtw but also provides the higher income/lower cost for each of CHP,e CHP,e FU @L the retailers/consumers relative to a single level model. ijtw ij CHP,h CHP,h CHP,h h =  - + - - -tw = 0 CHP,h ijtw CHP,h ijtw CHP,h ijtw The BL-HIA structure has presented an adequate frame- @P N ijtw i ij work for modelling both H-MGs reaction for a better par- (49) 10 FU 25 0.14 @L A B C A B C GB,h MCP SBP SSP GB GB,h h P P P P P P t t t t,e t,e t,e t,h t,h t,h = ijtw - +  -  -  = 0 (58) tw GB GB ijtw ijtw @P N ijtw ij 0.12 TES,h @L TES,h ijtw TES TES,h h =  +  -  - -  = 0 (59) ijtw tw TES ijtw TES ijtw @P 0.1 ijtw P ij TES,h TES,SOC @L TES TES = -   +  , TES ijtw ij(t-1)w ij ijtw @SOC ijtw 0.08 (60) TES,SOC 0TES 00TES + j + j = 0 t=1 t=24 ijw ijw ijtw 0.06 The complementary constraints obtained from the inequal- ity constraints of lower-level problem are in the form of fol- lowing relations: 0.04 CHP,e CHP,e CHP,e 0 6 P - P ?  > 0 (61) ij ijtw ijtw 0 0.02 CHP,e CHP,e CHP,e 00:00 05:00 10:00 15:00 20:00 0 6 P - P ?  > 0 (62) Time [h] ijtw ij ijtw CHP,h CHP,h CHP,h Figure 8: The profiles of the electrical and thermal loads of the H-MGs and 0 6 P - P ?  > 0 (63) ij ijtw ijtw electricity prices for buying and selling. CHP,h CHP,h CHP,h 0 6 P - P ?  > 0 (64) ijtw ij ijtw 20.00 WT WT,e 0 6 P ? > 0 (65) ijtw tw WT WT,e WT 15.00 0 6 P - P ? > 0 (66) ij ijtw ijtw ES ES ES,e 0 6 P - P ?  > 0 (67) ijtw ij STP ijtw 10.00 WT ES ES ES,e 0 6 P - P ?  > 0 (68) ij ijtw ijtw ES ES ES,SOC 5.00 0 6 SOC - P ?  > 0 (69) ijtw ij ijtw ES ES ES,SOC 0 6 SOC - SOC > 0 (70) ij ijtw ijtw 0.00 EB,h EB,h 12 3 45 6 78 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 0 6 P ?  > 0 (71) ijtw ijtw Figure 9: Shape of output power wave generated by WT and STP. EB,h EB,h EB,h 0 6 P - P ?  > 0 (72) ij ijtw ijtw EHP,h EHP,h @L 0 6 P ?  > 0 (73) WT WT,e WT,e e ijtw ijtw =  + - -  = 0 (50) ijtw tw tw WT tw @P ijtw EHP,h EHP,h EHP,h 0 6 P - P ?  > 0 (74) ij ijtw ijtw ES,e @L ijtw ES ES,e ES,e e TES TES TES,h =  +  -  - -  = 0 (51) 0 6 P - P ?  > 0 (75) ijtw tw tw ES ES ijtw ij tw ijtw @P ijtw P ij TES TES,h TES 0 6 P - P ?  > 0 (76) ij ijtw ijtw @L ES,e ES,e ES,SOC ES,SOC 0ES 00ES = - + - + j + j = 0 t=1 t=24 ijw ijw ES ijtw ij(t-1)w ijtw ijtw TES TES TES,SOC @SOC ijtw 0 6 SOC - P ?  > 0 (77) ijtw ij ijtw (52) TES @L TES TES,SOC EB,h EB,h EB,h h 0 6 SOC - SOC ?  > 0 (78) + +  -  -  = 0 (53) ij ijtw ijtw tw tw EB,h ijtw ijtw @P ijtw GB GB,h 0 6 P ?  > 0 (79) ijtw ijtw @L EB,e EB EB,h e = - -   +  = 0 (54) GB ij tw tw EB,e ijtw GB GB,h @P 0 6 P - P ?  > 0 (80) ij ijtw ijtw ijtw STP STP,h @L 0 6 P ?  > 0 (81) EHP,h EHP,h EHP,h EHP,h h ijtw ijtw =  + +  -  -  = 0 (55) ijtw ijtw ijtw tw EHP,h ijtw @P ijtw STP STP,h STP 0 6 P - P ?  > 0 (82) ij ijtw ijtw @L EHP,e EHP,h EHP e = - - COP  +  = 0 (56) e+ Grid,e tw EHP,e ijtw ijtw 0 6 P ?  > 0 (83) iktw @P iktw ijtw CHP,e Grid,e WT ES e+ 0 6  P + P + P - P ?  > 0 @L ijtw ijtw iktw STP,h ijtw iktw STP STP,h h =  +  -  -  = 0 (57) ijtw tw STP ijtw ijtw (84) @P ijtw Electrical and thermal power [kW] Electricity price [£/kWh] References References [1] Kohn W, Zabinsky ZB, Nerode A. A micro-grid distributed intel- ligent control and management system. IEEE Trans Smart Grid 2015;6(6):2964–74. [2] Liu Y, Yuen C, Hassan NU, Huang S, Yu R, Xie S. Electricity cost minimization for a microgrid with distributed energy resource under different information availability. IEEE Trans Ind Electron 2015;62(4):2571–83. [3] Li C, Yu X, Yu W, Chen G, Wang J. Efficient computation for sparse load shifting in demand side management. IEEE Trans Smart Grid 2017;8(1):250–61. [4] Zazo J, Zazo S, Macua SV. Robust worst-case analysis of demand-side management in smart grids. IEEE Trans Smart Grid 2017;8(2):662– [5] Ma J, Deng J, Song L, Han Z. Incentive mechanism for demand side management in smart grid using auction. IEEE Trans Smart Grid 2014;5(3):1379–88. [6] Pei W, Deng W, Shen Z, Qi Z. Operation of battery energy storage system using extensional information model based on IEC 61850 for micro-grids. IET Gener Transm Distrib 2016;10(4):849–61. [7] Zhang Y, Zhou J, Zheng Y, Xu Y. Control optimisation for pumped stor- age unit in micro-grid with wind power penetration using improved grey wolf optimiser. IET Gener Transm Distrib 2017;11(13):3246–56. [8] Liu G, Starke M, Xiao B, Tomsovic K. Robust optimisation-based micro- grid scheduling with islanding constraints. IET Gener Transm Distrib 2017;11(7):1820–28. [9] Liu C, Wang X, Wu X, Guo J. Economic scheduling model of micro- grid considering the lifetime of batteries. IET Gener Transm Distrib 2017;11(3):759–67. [10] Panwar LK, Konda SR, Verma A, Panigrahi BK, Kumar R. Operation window constrained strategic energy management of microgrid with electric vehicle and distributed resources. IET Gener Transm Distrib 2017;11(3):615–26. [11] Kinhekar N, Padhy NP, Li F, Gupta HO. Utility oriented demand side management using smart AC and micro DC grid cooperative. IEEE Trans Power Syst 2016;31(2):1151–60. [12] Basu AK, Bhattacharya A, Chowdhury S, Chowdhury SP. Planned scheduling for economic power sharing in a CHP-based micro-grid. IEEE Trans Power Syst 2012;27(1):30–38. [13] Lee S, Kwon B, Lee S. Joint energy management system of electric supply and demand in houses and buildings. IEEE Trans Power Syst 2014;29(6):2804–12. [14] Jayaweera D. Security enhancement with nodal criticality-based integration of strategic micro grids. IEEE Trans Power Syst 2015;30(1):337–45. [15] Wang P, Gao Z, Bertling L. Operational adequacy studies of power systems with wind farms and energy storages. IEEE Trans Power Syst 2012;27(4):2377–84. [16] Giusti A, Salani M, Caro GAD, Rizzoli AE, Gambardella LM. Restricted neighborhood communication improves decentralized demand-side load management. IEEE Trans Smart Grid 2014;5(1):92–101. [17] Ni J, Ai Q. Economic power transaction using coalitional game strategy in micro-grids. IET Gener Transm Distrib 2016;10(1):10–18. [18] Gabbar HA, Othman AM. Performance optimisation for novel green plug-energy economizer in micro-grids based on recent heuristic algo- rithm. IET Gener Transm Distrib 2016;10(3):678–87. [19] Marzband M, Parhizi N, Adabi J. Optimal energy management for stand-alone microgrids based on multi-period imperialist competition algorithm considering uncertainties: experimental validation. Int Trans Electr Energy Syst 2015;30(1):122–31. [20] Valinejad J, Marzband M, Akorede MF, Barforoshi T, Jovanovic M. Gen- eration expansion planning in electricity market considering uncer- tainty in load demand and presence of strategic gencos. Electric Power Systems Research 2017;152:92–104. [21] Valinejad J, Marzband M, Akorede MF, Barforoshi T, Jovanovi´ c M. Gen- eration expansion planning in electricity market considering uncer- tainty in load demand and presence of strategic GENCOs. Electr Power Syst Res 2017;152:92–104.

Journal

MathematicsarXiv (Cornell University)

Published: Oct 15, 2019

There are no references for this article.