ELECTRICITY DISTRIBUTION SYSTEM WITH DYNAMIC COOPERATIVE MICROGRIDS FOR REAL-TIME OPERATION

Abstract

System and methods for performing microgrids cooperation by optimal coalition formation in a distribution network are disclosed. A Microgrid Cooperation Module (MCM) is designed for utility EMS. MCM contains a coalition formation unit and an energy exchange decision unit. Furthermore, a communication protocol for energy exchange between two microgrids is designed. The coalition formation unit applies an innovative hierarchical coalition formation algorithm to provide optimal coalition for real time operation. The real time energy status of microgrids will be provided to coalition formation unit which will determine the coalitions (given a distance threshold) among microgrids to minimize the power loss. Energy exchange decision unit then determine actual energy transfer between pairs of microgrids within a coalition. Upon receiving the energy transfer information through a communication channel, the microgrids will start communicating and process energy transfer. The optimality of the formed coalitions is ensured by performing coalitional game theoretical analysis.

Claims

1. An electricity distribution system comprising several microgrids and one utility company where, each microgrid must be electrically connected with the utility company by medium voltage line and each microgrid may be connected with other microgrids by low voltage line; and the utility company and microgrids are interfaced with each other via communication network (e.g. wireless), further comprising: a utility energy management system (EMS), aims to reduce power loss within network as much as possible.

2. The electricity distribution system as set forth in claim 1, wherein the utility EMS comprising a Microgrid Cooperation Module (MCM) which contains two functional units 1) Coalition formation unit (CFU) for dynamically form coalitions of microgrids given their energy status which minimizes the power loss in network; 2) An energy exchange decision unit (EDU) for optimized exchange of energy among microgrids within a particular coalition given a distance threshold; a distribution network profile database that provides necessary spatial and temporal information of microgrids to both CFU and EDU.

3. The electricity distribution system as set forth in claim 2, wherein the microgrids comprise distributed energy resources and able to periodically declare energy status, that indicates the difference between total demand and total supply to the MCM via communication line.

4. The electricity distribution system as set forth in claim 2, wherein the CFU is able to process the received energy statues of microgrids and with the help of distribution network profile database can form optimal coalitions of microgrids; and an optimal coalition of microgrid can ensure Maximize intra-coalition energy exchange; Minimized distribution network power loss by utilizing low voltage lines among microgrids; Reduce energy burden on utility company.

5. The electricity distribution system as set forth in claim 2, wherein the EDU is able to decide the precise energy exchange profile within a coalition by enabling a predefined distance threshold which states the maximum allowable distance between two microgrids to initiate the energy exchange.

6. The electricity distribution system as set forth in claim 1, further comprising: a communication protocol for energy exchange between two microgrids.

7. The electricity distribution system as set forth in claim 4, wherein the CFU comprises a fast and scalable Hierarchical Priority based HR Coalition scheme to provide optimal coalition, in which the optimality property of the HR Coalition scheme is proved by Coalitional Game Theory.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0035] FIG. 1 is a perspective view of a communication interfaced (e.g. wireless) distribution system comprising several microgrids and utility system connected by medium and low voltage lines;

[0036] FIG. 2 shows the functional units in Microgrids Cooperative Module (MCM) which is located in utility energy management system;

[0037] FIG. 3 is an exemplary communication sequence diagram for energy exchange between two microgrids;

[0038] FIG. 4 is the process flow of Coalition Formation Unit (Proc. 1 in FIG. 2);

[0039] FIG. 5 is the process flow of Energy Exchange Decision Unit (Proc. 2 in FIG. 3);

[0040] FIG. 6 is an exemplary figure showing increase in utility after Proc. 1 and Proc. 2 for a 10 microgrids distribution system;

[0041] FIG. 7 is a figure showing increase in Shapley Value after Proc. 1 and Proc. 2 for a 10 microgrids distribution system using the same system data as of FIG. 6;

[0042] FIG. 8 is an exemplary figure showing utility values of coalitions in a 100 microgrids distribution system;

[0043] FIG. 9 is an exemplary power loss reduction comparison for a 200 microgrids distribution system;

[0044] FIG. 10 is an exemplary power loss reduction comparison for a 500 microgrids distribution system;

[0045] FIG. 11 is a summarized power loss phenomena for different scaled distribution system;

[0046] FIG. 12 is the pattern of average execution time (AET) with the size of the distribution system (from exemplary 5 microgrids to 500 microgrids system).

DESCRIPTION OF EMBODIMENTS

[0047] Hereinafter, some exemplary embodiments of the present invention including the models, methods and numerical test results are described in details with reference to the accompanying drawings.

Modeling Distribution System for the Invention

[0048] Generally, microgrids are operated in grid-connected mode. That is, when the microgrid requires energy to meet its internal demand, the utility grid provides the additional energy. At the same way, whenever a microgrid has surplus of energy, it will sell the energy to utility grid. Therefore, the traditional distribution system architecture is composed with bi-directional energy and electricity communication between a microgrid and the utility company/grid.

[0049] The high level exemplary distribution system diagram is shown in FIG. 1. The exemplary system consists 5 microgrids each of which is connected to the utility grid via a medium voltage line. The real time energy status (in a form of signal) from each microgrid to utility energy management system (EMS) will be conducted through communication line. The utility EMS can, at the same time, send appropriate signals to microgrids via communication line. A subset of microgrids can be connected with each other via low voltage line. Depending on the geographical location and spatial configuration, these low voltage lines are established. These physically connected microgrids can establish communication chat via communication infrastructure (the details will be described in FIG. 3).

[0050] Differences with model described in Prior Art 2:

[0051] The system depicted in the Prior Art 2 also contains several microgrids connected with one utility company via electric lines with different voltage level. However, the communication infrastructure, network architecture and system model is quite different than that of the invented model. The significant differences are:

[0052] 1. In Prior Art 2, the microgrids operate in distributed fashion, where every microgrid has to report its energy status as well as spatial information to every other microgrid in the network. Such infrastructure is highly vulnerable to security leak and yields reliability issue. In the present invention, the microgrids only report their energy status to the utility EMS (which requires minimum communication and is considered sufficiently reliable and secure since microgrids are sharing bare minimum information to utility company).

[0053] 2. In Prior Art 2, coalition formation intelligence needs to be installed in every microgrid (possibly, in smart meter or other microgrid energy unit; it is not clear from the description of Prior Art 2). In the present invention, the intelligence of coalition formation and energy exchange is located centrally to utility EMS (more precisely, in MCM).

Detailed Description of the Invented Functional Units in MCM

[0054] The designed functional unit which hosts the coalition formation and energy exchange management methods in utility EMS is detailed in FIG. 2. The utility EMS is typically responsible for optimized energy operation of the distributed system under that particular utility company. The utility EMS periodically receives energy status information from microgrids via communication line (as shown in FIG. 2). The Microgrid Cooperation Module (MCM) contains two functional units, 1) Coalition Formation Unit (CFU), and 2) Energy exchange Decision Unit (EDU). The distribution network profile will provide spatial information to both of these subunits for performing respective tasks. Such information will be kept in a localized database which will be updated when necessary. Upon receiving energy status information from microgrids, the coalition formation subunit will form the optimal coalition of microgrids which will minimize the energy exchange with utility company, maximize the inter-microgrid energy transfer. After achieving optimal coalition, the EDU will determine the potential optimal energy exchange energy exchange among microgrids. This unit will produce an Energy Transfer matrix and break it down to microgrid level. Associated signal and table are then sent to corresponding microgrid via communication network. The signal contains either send energy or receive energy followed by a table of energy amount to perspective other microgrids (FIG. 2).

Coalition Formation Unit (CFU)

[0055] The detailed process flow of CFU is shown in FIG. 4. The process basically tries to pair up two existing coalitions, one with the highest aggregated load and another with highest aggregated supply, (if their centroids are locating within a predefined proximity) until no possible updates in coalitions can possible. The invented coalition mechanism can be named as Hierarchical Priority based coalition formation or HR Coalition. Apparently, i is a load microgrid iff E.sub.i<0 or a provider microgrid iff E.sub.i>0. [0056] Step 1: The process stats with initialization step where all participating microgrids are treated as individual coalition and their centroids are their actual location. [0057] Step 2: Next step is to group the coalitions into two sets, 1) the load group containing the microgrids with negative energy status (LcG) and 2) the provider group containing positive energy status (PcG). The 3rd and 4th steps will be in repeat as long as there is a change in coalitions. [0058] Step 3: The 3rd step orders LcG according to high to low of the load demand (OLG) and PcG according to high to low of supply (OPG). [0059] Step 4: The 4th step contains several sub-steps. For each coalition C.sub.1 in OLG [0060] C.sub.p is 1st coalition in OPG whose central is located within the proximity of C.sub.1. If no such C.sub.p is found in OPG, make no change and scan the next C.sub.1 on the list. [0061] Integrate the coalitions C.sub.1 and C.sub.p into C.sub.1 [0062] Remove C.sub.p from PcG [0063] Update the energy status and centroid information of C.sub.1 [0064] The process goes back to Step 3 if there is any change in any of the coalitions [0065] Step 5: Check whether there is still a change in coalitions. If so (i.e. some coalitions are updated), the process goes back to Step 2. [0066] Step 6: Terminates with the final coalition C.sub.f=C.sub.1C.sub.p

Energy Exchange Decision Unit (EDU)

[0067] The CFU then sends the final microgrid coalitions C.sub.f to the EDU in order to determine energy transfer matrix. The process in CFU can be interpreted as a Hierarchical priority based intelligent coalition scheme (HR Coalition). The detailed process flow of EDU is shown in FIG. 5. In each coalition, EDU basically forms a 1 to 1 mapping between a load microgrid and a supply microgrid who are locating within a proximity threshold. Before, the mapping process starts, EDU orders the microgrids according to their energy status [0068] The load microgrids are ordered according to higher load to lower load [0069] The supplier microgrids are ordered according to higher supply amount to lower supply amount.

[0070] Such technique will ensure maximum possible energy transfer within microgrids.

Algorithmic Complexity of Optimal Coalition Formation and Comparison

[0071] Optimal coalition formation of microgrids will ensure minimized power loss and well as maximized inter-microgrid energy exchange. Forming such coalition, however, is computationally intensive as the number of microgrids grows higher and inherently complex given a distribution network profile. The conventional mathematical optimization method (such as Linear programming) can ensure the optimality provided the correct mathematical model is formulated. However, the complexity of such method is exponential with the number of microgrid. To be more precise, since the method has to check all possible combination, the algorithmic complexity is O(2.sup.|N|). Thus Optimal Coalition formation is an NP-Complete problem. Therefore, it is computationally almost impossible to perform optimal coalition formation using mathematical optimization methods when the number of microgrids exceeds a particular threshold. Moreover, as pointed before, the game theoretic merge/split operation is an NP-hard problem. Thus, it is impossible to solve the operation in a polynomial time, if the number of microgrids is higher than a specific number. Applying some heuristics and assumption (as done in Prior Art 2), the complexity can be brought down to a tolerable range. However, even the reduced complexity of merge and split is not sufficient enough to be applicable in a real-time operation with a very high number of microgrids. On the other hand, the invented coalition formation algorithm (namely HR Coalition) is a priority based hierarchical scheme, which tries to form coalition based on the energy status of the microgrid. The computational complexity of HR Coalition, therefore, is O(|N|.sup.2).

[0072] The communication complexity of merge/split operation used in Prior Art 2 is O(|N|.sup.3) since every microgrid has to communicate with every other microgrid in order to receive the energy and network information and again in transfer of energy. The present invention, on the other hand, has a worst case communication complexity of O(|N|.sup.2). Because, after deciding the energy transfer between the microgrids, each microgrid has to communicate its corresponding microgrid only one time.

Description of the Invented Protocols for Inter-Microgrid Communication

[0073] An exemplary simplified communication sequence diagram for processing energy transaction between two microgrids is shown in FIG. 3. The communication protocol is assumed to be synchronous. Both of the communicating microgrids first receive the signal and energy transaction table from utility EMS. The load microgrid (MG 1, in FIG. 3) initiates the communication while the provider microgrid (MG 4, in FIG. 3) is waiting for any load microgrid to send initiate signal. The initiate signal also checks whether the communication link is active or not. After receiving the signal, MG 4 acknowledges. MG 1 then sends a request signal with the energy amount it requires from MG 4 (which was initially sent by the utility EMS). Upon receiving the request signal, MG 4 checks the energy amount with the same one it receives from utility EMS. Ideally, these two amounts are supposed to be same. MG 4, however, can decline the request if the energy amount signal it receives from MG 1 is not as same as the energy amount signal from utility EMS. In case, these signals are matched, MG 4 sends a prepare signal to MG 1, stating MG 1 to prepare necessary hardware and logistics (the substation equipment, necessary switches, circuit breakers, etc.) require to receive the power from MG 4. After preparation, MG 1 sends the ready signal to MG 4 stating that it is ready to receive energy (18.10 kWh, in this case). The provider microgrid (MG 4) then transfer power to load microgrid (MG 1) using the low voltage line connected between them. The communication terminates after MG 4 sends a transfer-status saying whether the transfer is a success or a failure.

Numerical Simulation and Analyses

[0074] In order to compare the effectiveness of the method in CFU, an equivalent distance based clustering coalition scheme is implemented.

10 Microgrids and 100 Microgrids Cases

[0075] An exemplary case of 10 microgrids in a distribution system is considered. These microgrids are assumed to be scattered randomly over a 5 square kilometer area. The utility grid is assumed to be located at the center of the area. The intra coalition distance threshold is set to be 2.5 km. FIG. 6 shows utility of each microgrid after forming coalition. It is evident from the figure that, forming coalitions using HR coalition scheme effectively increases (or at least kept same) the utilities of microgrids. FIG. 7 describes the Shapley value phenomena. The microgrids are, once again, gained their marginal contribution by joining the coalition formed by HR Coalition scheme invented as CFU. Another exemplary case of 100 microgrids spaced over 10 km with an internal coalition centroid distance threshold of 1.5 km is considered. The number of coalitions formed is 38. The utility for each coalition is printed in FIG. 8.

Power Loss Phenomena

[0076] The power loss reduction phenomena realized by the invented CFU and EDU are shown in FIG. 9, FIG. 10 and FIG. 11 for some exemplary systems. FIG. 9 describes the reduction of power loss achieved by the 10 microgrid case and a comparison with equivalent DO Coalition scheme. The result is conducted by running both of these methods for 100 times. The HR Coalition scheme (in CFU) effectively reduces the loss of power within the network in almost most of the (86%) times. The average loss reduction is 26%. FIG. 10 shows the loss reduction phenomena when the number of microgrids is 500. It is very interesting to see, in this case, the loss is reduced (significantly) in all of the runs (100%). Moreover, the average loss reduction is also improved to 77%. Therefore, the coalitions formed by CFU scheme are scalable and performs better with higher scaled system. To summarize the loss reduction phenomena, FIG. 11 is presented. This figure portrays the average percentage of power loss reduced after applying HR Coalition scheme for different sized system. As the system size grows, the invented HR coalition scheme reduces the power loss in the network. Another noteworthy point needs to be mention is that the average loss per microgrid is also reduced with the system size. Precisely speaking, the average loss per microgrid in 10-microgrid case is 0.20 kW while the same in 500-MG case is 0.14 kW while performing coalition formation through CFU for both of the cases.

Average Execution Time Pattern

[0077] The pattern of average execution time (AET, in seconds) of forming coalitions is shown in FIG. 12. The processes in CFU and EDU are conducted 100 times each over a hypothetical distribution network of 5 microgrids to 500 microgrids (with a step size of 5). FIG. 12 suggests that the AET has a quadratic relationship with number of microgrids. Compared with the conventional coalition formation method (which yields an exponential relationship between AET and number of microgrids), the invented method is supremely fast. Hence, the invented method can be effectively applicable in real time.