A linear optimization method for real-time energy dispatching of isolated microgrids

Abstract

The present invention discloses a linear optimization method for real-time dispatch of energy in an isolated microgrid, including designing an isolated microgrid structure based on a multi-agent system, designing a real-time dispatch optimization method for the isolated microgrid, and establishing a real-time dispatch linear model for the isolated microgrid. The isolated microgrid increases the absorption resistance to absorb the excess power to ensure real-time power balance; by establishing a linear model of a battery, a micro gas turbine and an absorption resistance, the isolated microgrid is transformed into a simple linear programming model with a small amount of calculation, and then the intelligent agent is used to collect data information such as the output plan of the battery and micro gas turbine scheduled before the day, the deviation value of renewable energy and load between the day-ahead dispatch and the real-time dispatch, and the real-time dispatch plan of energy is obtained by optimization through a linear programming solution algorithm. The present invention can improve the speed and ability of the isolated microgrid to handle the uncertainty of load and renewable energy in real time, and improve the reliability of the stable operation of the isolated microgrid.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart grids and relates to a linear optimization method for real-time energy scheduling of isolated island microgrids (MG). Background Art

[0002] In the day-ahead energy scheduling optimization of isolated MGs, due to the many uncertainties in load and renewable energy sources (RES), such as user electricity consumption and weather changes, there are varying degrees of deviation between the predicted load and RES data and the actual data. Therefore, the output plan for each unit developed by the day-ahead energy scheduling optimization of isolated MGs may not necessarily match the actual energy scheduling situation.

[0003] To this end, on the basis of day-ahead scheduling optimization, it is necessary to make certain adjustments to the output plan of each unit in real-time scheduling optimization to cope with the changes in actual Load and RES, so that the island MG can meet the real-time power balance. At the same time, the real-time scheduling optimization cost of the island MG is minimized. However, real-time scheduling optimization requires the MG system to be able to quickly make real-time dynamic adjustments when responding to real-time changes in Load and RES. In real-time scheduling optimization, Load and RES are regarded as uncontrollable units due to their uncertainty factors. Therefore, real-time scheduling optimization can only be achieved by making real-time dynamic adjustments to other units (such as battery energy storage system (BESS), micro turbine (MT), etc.).

[0004] However, in the traditional island MG real-time scheduling optimization model, there are many complex models with nonlinear relationships, which leads to a decrease in the solution speed and the quality of the optimal solution in real-time scheduling optimization, thereby affecting the speed and ability of the island MG to handle load and RES uncertainties in real time. Summary of the Invention

[0005] The present invention aims to overcome the shortcomings of the prior art and provide a linear optimization method for real-time energy scheduling of isolated MGs. By establishing a linear model for real-time scheduling of isolated MGs, real-time scheduling optimization is achieved, thereby maximizing the speed of solving the real-time energy scheduling optimization of isolated MGs and obtaining a unique high-quality solution. This improves the speed and ability of the isolated MGs to handle load and RES uncertainties in real time, thereby enhancing the reliability of the stable operation of the isolated MGs.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions.

[0007] A linear optimization method for real-time scheduling of energy in an isolated island microgrid of the present invention comprises the following steps:

[0008] Step 1: Design an island microgrid structure based on a multi-agent system: its structure includes renewable energy, micro gas turbines, loads, and absorption resistors; design the connection relationship between the microgrid, renewable energy, micro gas turbines, loads, absorption resistors, and agents; and explain the working relationship between each agent;

[0009] Step 2: Design a real-time scheduling optimization method for the isolated microgrid. This includes collecting the deviation between the day-ahead load and the real-time load, and the deviation between the day-ahead renewable energy and the real-time renewable energy; sending the day-ahead battery charge / discharge power plan and the day-ahead micro-gas turbine power output plan to the isolated microgrid energy real-time scheduling linear model; executing the real-time energy scheduling optimization to obtain the actual output plan of each unit and feeding it back to each unit;

[0010] Step 3: Establish a real-time dispatch linear model for the island microgrid, including the linear model of the battery, the linear model of the micro gas turbine, the linear model of the absorption resistor, the linear objective function of the real-time dispatch of the island microgrid energy and its constraints.

[0011] Furthermore, the design of step 1 is based on the island microgrid structure of the multi-agent system, including the following processes:

[0012] The island microgrid structure based on the multi-agent system includes renewable energy, batteries, micro gas turbines, loads and absorption resistors; the renewable energy, batteries, micro gas turbines, loads and renewable energy are interconnected through power transmission lines; the renewable energy, batteries, micro gas turbines, loads and absorption resistors are respectively connected to corresponding intelligent agents, wherein the intelligent agents of renewable energy, batteries, micro gas turbines, loads and absorption resistors are respectively connected to the microgrid intelligent agent; the intelligent agent connected to the renewable energy intelligent agent is used to collect the power output information of the renewable energy in real time; the intelligent agent connected to the battery is used to collect the battery's state of charge, upper and lower limits of charge and discharge power, and energy storage capacity information in real time, and send charge / discharge power instructions to the battery; and the intelligent agent connected to the micro The intelligent agent connected to the gas turbine is used to collect the power output information and power output constraints of the micro gas turbine in real time, and send power output instructions to the micro gas turbine; the intelligent agent connected to the load is used to collect the load value at each moment in real time; the intelligent agent connected to the absorption resistor is used to collect the number of absorption resistors and the amount of constraint information on the absorption of excess power in real time, and send the amount of excess power to be consumed to the absorption resistor; the microgrid intelligent agent is responsible for real-time energy scheduling optimization; the intelligent agents connected to renewable energy, batteries, micro gas turbines, loads and absorption resistors need to send all collected data information to the microgrid intelligent agent; the energy scheduling instructions optimized by the microgrid intelligent agent are sent to the intelligent agents connected to the batteries, micro gas turbines and absorption resistors.

[0013] Furthermore, the specific process of the step 2 for designing the real-time dispatch optimization method for the isolated island microgrid is as follows:

[0014] First, the load agent collects the actual load and the day-ahead load and calculates their error value; the renewable energy agent collects the actual renewable energy and the day-ahead renewable energy and calculates their error value; then the load agent and the renewable energy agent send the obtained error values ​​to the microgrid agent respectively; after receiving the error values ​​from the load agent and the renewable energy agent, the microgrid agent calculates the total error value, namely: [actual load - day-ahead load] - [actual renewable energy - day-ahead renewable energy], and sends the total error value to the real-time scheduling linear model of the isolated microgrid; at the same time, the battery agent and the micro gas turbine agent respectively send the battery in the day-ahead scheduling optimization. The charging / discharging power plan and the power output plan of the micro gas turbine are sent to the real-time scheduling linear model of the isolated microgrid; then, the microgrid intelligent agent uses a linear programming solver to perform real-time energy scheduling optimization on the real-time scheduling linear model to obtain the actual output plan of each unit; the microgrid intelligent agent sends the actual output plan of each unit to the battery intelligent agent, micro gas turbine intelligent agent, and absorption resistor intelligent agent respectively; the battery intelligent agent, micro gas turbine intelligent agent, and absorption resistor intelligent agent send the corresponding actual output plan instructions to the battery, micro gas turbine, and absorption resistor; finally, the battery, micro gas turbine, and absorption resistor operate according to the actual output plan.

[0015] Furthermore, the algorithm of the linear programming solver in step 2 includes the following steps:

[0016] 1) Input: (1) Battery parameter input: maximum energy storage capacity E, each scheduling interval Δt, maximum charge / discharge power Unit operating cost r1, r2, r3, initial state of charge value SOC initial ;

[0017] (2) Micro gas turbine parameter input: maintenance factor K MT , fuel unit price λ, maximum power output

[0018] (3) Absorption resistance parameter input: unit cost of power loss τ of the island microgrid, maximum amount of excess power to be absorbed

[0019] (4) Other parameter input: deviation value and Battery charging / discharging power plan P in day-ahead dispatch optimization t BESS and microturbine output plan P t MT ;

[0020] 2) Real variable setting: Battery real-time dynamic adjustment of power P taBESS , Micro gas turbine real-time dynamic adjustment of power P t aMT , the amount of excess power absorbed by the absorption resistor P t CR ;

[0021] 3) Constraint setting: Formula (6): Formula (8): 0≤SOC t ≤1, formula (9): Formula (11): Formula (12):

[0022]

[0023] 4) Objective function setting:

[0024] 5) Optimization solution:

[0025] 6) Output: The battery dynamically adjusts the optimal power value in real time Real-time dynamic adjustment of the power optimal value of micro gas turbine The optimal value of the number of resistors that can absorb excess power

[0026] Furthermore, the step 3 establishes a real-time scheduling linear model for the isolated microgrid, including: the real-time scheduling linear model for the isolated microgrid includes a linear model of the battery, a linear model of the micro gas turbine, a linear model of the absorption resistor, a real-time scheduling linear objective function and constraints of the isolated microgrid, and the steps are as follows:

[0027] Step 3.1. Linear model of battery

[0028] The operating cost of a battery is related to the charge / discharge power and the state of charge. The unit operating cost of each operating area of ​​the battery can be set separately for the discharge and charge processes:

[0029] (1) When the battery is discharging, if 0.7<SOC t ≤1, the unit operating cost of discharge is r1; if 0.7≤SOC t ≤0.3, the unit operating cost of discharge is r2; if 0.3<SOC t ≤0, the unit operating cost of discharge is r3; where r1<r2<r3 and r1, r2, r3 are generally taken as 0.05, 0.1, 0.25;

[0030] (2) When the battery is charging, if 0.7<SOC t≤1, the unit operating cost of charging is r3; if 0.7≤SOC t ≤0.3, the unit operating cost of charging is r2; if 0.3<SOC t ≤0, the unit operating cost of charging is r1;

[0031] It can be seen from this that the battery operation cost function is a piecewise function. In order to improve the calculation speed of the battery real-time scheduling operation cost, the following linear model is established:

[0032]

[0033]

[0034] Where, is the real-time dispatching operation cost of the battery, is the average unit operating cost of the battery, SOC t is the state of charge of the battery at time t, E is the maximum energy storage capacity of the battery, Δt is the interval time of each dispatch, P t RBESS Real-time scheduling of charge / discharge power for batteries;

[0035] Step 3.2: Linear model of microturbine

[0036] The operating cost of a micro gas turbine includes maintenance cost and fuel cost. Fuel cost is related to the unit price of fuel, the lower calorific value of natural gas, operating efficiency, and power output. The operating efficiency of a micro gas turbine is a cubic function of power output. The operating efficiency of a micro gas turbine must be calculated first before the operating cost of the micro gas turbine can be calculated. To improve the calculation speed of the real-time scheduling operating cost of a micro gas turbine, the following linear model is established:

[0037]

[0038] Where, is the real-time dispatching cost of the micro gas turbine, K MT and λ are the maintenance factor and fuel price of the micro gas turbine respectively, P t RMT Real-time dispatch of power output for micro-turbines;

[0039] Step 3.3: Linear model of the absorber resistor

[0040] The excess power absorbed by the absorption resistor in the isolated island microgrid is converted into power loss cost. The real-time dispatching operation cost of the absorption resistor can be defined as the following linear model:

[0041]

[0042] Where, is the real-time dispatching operation cost of the absorption resistor, τ is the unit cost of power loss in the isolated microgrid, P t CR To absorb the amount of excess power consumed by the resistor in real-time scheduling;

[0043] Step 3.4: Real-time dispatch linear objective function of island microgrid

[0044] The real-time dispatching and operating costs of the isolated microgrid, including the real-time dispatching and operating costs of batteries, micro gas turbines, and absorption resistors, are calculated by establishing the following objective function:

[0045]

[0046] Where, The operating cost of real-time energy dispatch for isolated microgrids is and They are all linear models, so Equation (5) is called the real-time scheduling linear objective function of the island microgrid;

[0047] Step 3.5: Real-time dispatch linear constraints for island microgrids

[0048] In real-time scheduling optimization, batteries, micro gas turbines, and absorption resistors must not only meet their own constraints but also real-time power balance constraints. Therefore, the real-time scheduling linear model of the island microgrid must meet the following constraints:

[0049] (1) Battery charge and discharge power constraints and state of charge constraints

[0050]

[0051] P t RBESS =P t aBESS +P t BESS (7)

[0052]

[0053] Where, and are the maximum charging power and maximum discharging power of the battery, respectively, where P t RBESS A positive value indicates discharge, P t aBESS and P t BESS The battery is used to dynamically adjust the power in real-time scheduling and the charge / discharge power plan in day-ahead scheduling. SOCinitial is the initial value of the battery's state of charge;

[0054] (2) Power output constraints of micro gas turbines

[0055]

[0056] P t RMT =P t aMT +P t MT (10)

[0057] Where, is the maximum power output of the micro gas turbine, P t aMT and P t MT The micro gas turbine dynamically adjusts power in real-time scheduling and plans power output in day-ahead scheduling.

[0058] (3) Quantity constraints on the amount of excess power absorbed by the absorption resistor

[0059]

[0060] Where, is the maximum amount of excess power that can be absorbed by the absorption resistor;

[0061] (4) Real-time power balance constraints for real-time dispatch optimization of isolated microgrids

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] Where, and are the deviations between the actual load and renewable energy in real-time scheduling and the load and renewable energy output plan in day-ahead scheduling optimization, P t ALoad and P t PLoad They are the actual load in real-time scheduling and the load output plan in day-ahead scheduling optimization, P t ARES and P t PRESThey are the actual renewable energy in real-time scheduling and the renewable energy output plan in day-ahead scheduling optimization, P t Load and P t un are the number of forecast load and unloaded load in day-ahead dispatch optimization, P t RES and P t abRES They are the amount of predicted renewable energy and abandoned renewable energy in day-ahead dispatch optimization, respectively.

[0068] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0069] 1. The island MG structure with consumption resistors (CR) adopted in this invention facilitates the island MG to utilize the CR to consume excess power in the system in real time, improving the island MG's ability to handle the uncertainty between load and RES in real time and its stable operation reliability.

[0070] 2. The real-time scheduling optimization method for island MGs designed in this invention makes small, real-time dynamic adjustments to each unit's output plan based on the island MG's day-ahead scheduling plan. This helps island MGs quickly handle uncertain fluctuations in load and resource availability during real-time scheduling.

[0071] 3. The linear model for real-time scheduling of isolated MGs established in this invention is linear, simple, and computationally lightweight. Therefore, it helps improve the speed of real-time energy scheduling optimization for isolated MGs, facilitates obtaining a unique, high-quality solution, and further enhances the speed and ability of isolated MGs to handle load and RES uncertainties in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a structural diagram of an isolated MG based on a multi-agent system (MAS) according to an embodiment of the present invention.

[0073] Figure 2 This is a design diagram of a real-time scheduling optimization method for an isolated MG according to an embodiment of the present invention.

[0074] Figure 3 This is a flow chart of a linear optimization method for real-time scheduling of isolated island MG energy according to an embodiment of the present invention.

[0075] FIG4 is a simulation diagram of a case of real-time scheduling of energy for isolated MGs according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] The present invention discloses a linear optimization method for real-time scheduling of energy in an isolated microgrid, including the design of an isolated microgrid structure based on a multi-agent system, the design of a real-time scheduling optimization method for the isolated microgrid, and the establishment of a real-time scheduling linear model. The isolated microgrid increases a dissipation resistor to dissipate excess power to ensure real-time power balance; by establishing a linear model of a battery, a micro gas turbine, and a dissipation resistor, the real-time scheduling objective function and constraint conditions of the isolated microgrid with the minimum operating cost are transformed into a simple linear programming model with low computational complexity; then, the intelligent agent is used to collect data information such as the output plan of the battery and micro gas turbine for day-ahead scheduling, the deviation value of renewable energy and load between day-ahead scheduling and real-time scheduling, and the real-time energy scheduling plan is obtained by optimization and solution through a linear programming solution algorithm. The present invention can improve the speed and ability of the isolated microgrid to handle the uncertainty of load and renewable energy in real time, and enhance the reliability of the stable operation of the isolated microgrid.

[0077] The description of the technical solution of the present invention involves the following English abbreviations:

[0078] MG: microgrid. Load: load. RES: renewable energy source. BESS: battery. MT: microturbine. MAS: multi-agent system. CR: absorption resistor. Agent: agent. SOC: state of charge. MGA: microgrid agent. RA: renewable energy source agent. LA: load agent. BA: battery agent. MA: microturbine agent. CRA: absorption resistor agent.

[0079] The present invention will be further described in detail below with reference to the accompanying drawings.

[0080] Figure 3 FIG1 is a flow chart of a linear optimization method for real-time scheduling of isolated island MG energy according to an embodiment of the present invention. Figure 3 As shown, a linear optimization method for real-time scheduling of isolated island MG energy of the present invention includes the following steps:

[0081] Step 1. Structure diagram of isolated MG (microgrid) based on MAS (multi-agent system): The isolated MG (microgrid) structure includes RES (renewable energy), MT (micro gas turbine), Load and CR (absorptive resistor), as well as the connection relationship between MG (microgrid), RES (renewable energy), MT (micro gas turbine), Load and CR (absorptive resistor) and the agent, and the working principle between the agents is explained in detail.

[0082] Step 2. Design diagram of the real-time scheduling optimization method for the isolated MG (microgrid): This method includes collecting the deviation value between the day-ahead Load and the real-time Load, as well as the deviation value between the day-ahead RES (renewable energy) and the real-time RES (renewable energy), sending the day-ahead BESS (battery) charge / discharge power plan and the day-ahead MT (micro gas turbine) power output plan to the isolated MG (microgrid) energy real-time scheduling linear model, performing energy real-time scheduling optimization, obtaining the actual output plan of each unit, and feeding back the actual output plan to each unit.

[0083] Step 3: Establishment of a linear model for real-time dispatch of isolated MG: This model includes the linear model of BESS (battery), the linear model of MT (micro gas turbine), the linear model of CR (consumption resistance), the linear objective function of real-time dispatch of energy of isolated MG (microgrid) and its constraints.

[0084] The specific process of step 1 is as follows: the island MG structure includes RES, BESS, MT, Load, and CR, wherein RES, BESS, MT, Load, and RES are interconnected via power transmission lines; RES, BESS, MT, Load, and CR are respectively connected to corresponding Agents, wherein the Agents of RES, BESS, MT, Load, and CR are respectively connected to the MG Agent (MGA). The Agent connected to RA is used to collect real-time power output information of RES; the Agent connected to BESS is used to collect real-time information such as BESS state of charge, upper and lower limits of charge and discharge power, and energy storage capacity, and send charge / discharge power instructions to BESS; the Agent connected to MT is used to collect real-time power output information and power output constraints of MT, and send power output instructions to MT; the Agent connected to Load is used to collect real-time Load values ​​at each moment; the Agent connected to CR is used to collect real-time information such as the number of CRs and the number of constraints on excess power absorption, and send the amount of excess power to be consumed to CR; MGA is mainly responsible for real-time energy scheduling optimization. The agents connected to the RES, BESS, MT, Load, and CR must transmit all collected data to the MGA. Similarly, the energy dispatch instructions optimized by the MGA must be sent to the agents connected to the BESS, MT, and CR. The CR is introduced to account for the situation where, during real-time energy dispatch, an island MG experiences excess power that cannot be fully absorbed by the BESS. Therefore, this structure improves the island MG's ability to handle load and RES uncertainties during real-time energy dispatch, and enhances the reliability of its stable operation.

[0085] Figure 1This is an islanded MG architecture based on a MAS. The islanded MG architecture consists of the RES, BESS, MT, Load, and CR. The MAS, in turn, consists of local agents and MGAs. Local agents include the RES Agent (RA), Load Agent (LA), BESS Agent (BA), MT Agent (MA), and CR Agent (CRA). They are responsible for collecting and classifying data from each unit (i.e., RES, Load, BESS, MT, and CR) and issuing power commands to each unit. The local agent communicates with the MGA and uploads relevant data. The MGA performs MG-level optimization based on the data received from the local agent and sends the power change commands for each unit back to the local agent. MG-level optimization includes day-ahead scheduling optimization and real-time scheduling optimization. The CR is only used for real-time scheduling optimization. Data can be transmitted between each local agent and the corresponding unit, and between each local agent and the MGA, via information transmission lines. Power can be transmitted between the RES, BESS, MT, Load, and CR via power transmission lines.

[0086] The specific process of step 2 is as follows: first, the LA collects the actual load and the day-ahead load and calculates the error between them (actual load - day-ahead load); the RA collects the actual RES and the day-ahead RES and calculates the error between them (actual RES - day-ahead RES); then, the LA and RA respectively send the obtained error values ​​to the MGA; after receiving the error values ​​provided by the LA and RA, the MGA calculates the total error value ([actual load - day-ahead load] - [actual RES - day-ahead RES]) and sends the total error value to the real-time scheduling linear model of the isolated MG; simultaneously, the BA and MA respectively send the charging / discharging power plan of the BESS and the power output plan of the MT in the day-ahead scheduling optimization to the real-time scheduling linear model of the isolated MG; then, the MGA uses a linear programming solver to perform real-time energy scheduling optimization on the real-time scheduling linear model to obtain the actual output plan of each unit; the MGA sends the actual output plan of each unit to the BA, MA, and CRA respectively; the BA, MA, and CRA send the corresponding actual output plan instructions to the BESS, MT, and CR; finally, the BESS, MT, and CR operate according to the actual output plan.

[0087] Figure 2 The diagram shows the design of the real-time scheduling optimization method for isolated MG. First, LA and RA collect the actual Load (P t ALoad ) and actual RES(P t ARES) data, and calculate the Load output plan (P t PLoad ) and RES output plan (P t PRES ) between the deviation values ​​( and ). Then, LA and RA will and Send to MGA, MGA will calculate the total error value The total error value is sent to the real-time scheduling linear model of the isolated island MG. Meanwhile, BA and MA send the BESS charging / discharging power plan and MT power output plan at time t in the day-ahead scheduling optimization to the real-time scheduling linear model of the isolated island MG in MGA respectively. MGA performs energy real-time scheduling optimization on the real-time scheduling linear model of the isolated island MG by using a linear programming solver, and obtains the actual output plans (P t RBESS 、P t RMT and P t CR ); the MGA sends the actual output plans of the BESS, MT, and CR to the BA, MA, and CRA, respectively; the BA, MA, and CRA send power output plan instructions to the BESS, MT, and CR; finally, the BESS and MT execute the corresponding outputs according to the power instructions, and the CR connects the corresponding amount of absorption resistors according to the power instructions. This completes the real-time energy scheduling of the island MG at time t. The linear programming solver algorithm steps include:

[0088] 1) Input: (1) BESS parameter input: maximum energy storage capacity E, each scheduling interval Δt, maximum charge / discharge power

[0089] Unit operating cost r1, r2, r3, initial state of charge value SOC initial ;

[0090] (2) MT parameter input: maintenance coefficient K MT , fuel unit price λ, maximum power output

[0091] (3) CR parameter input: unit cost of power loss of isolated MG, τ, and maximum number of excess power to be absorbed.

[0092] (4) Other parameter input: deviation value and BESS charging / discharging power plan P in day-ahead dispatch optimization t BESSand MT output plan P t MT ;

[0093] 2) Real variable setting: BESS dynamically adjusts power P in real time t aBESS , MT real-time dynamic adjustment of power P t aMT 、The amount of excess power absorbed by CR P t CR ;

[0094] 3) Constraint setting: Formula (6): Formula (8): 0≤SOC t ≤1, formula (9): Formula (11): Formula (12):

[0095]

[0096] 4) Objective function setting:

[0097] 5) Optimization solution:

[0098] 6) Output: BESS dynamically adjusts the optimal power value in real time MT dynamically adjusts the optimal power value in real time The optimal number of CRs to absorb excess power

[0099] The specific process of step 3 is as follows: the real-time scheduling linear model of the isolated island MG includes the linear model of the BESS, the linear model of the MT, the linear model of the CR, the real-time scheduling linear objective function of the isolated island MG and the constraints. The steps are as follows:

[0100] Step 3.1: Linear model of BESS

[0101] The operating cost of a BESS is related to the charge / discharge power and the state of charge (SOC). The unit operating cost of each working area of ​​the BESS is set according to the charge and discharge process, and can be set separately for the discharge and charge processes:

[0102] (1) When the BESS is in discharge, if 0.7<SOC t ≤1, the unit operating cost of discharge is r1; if 0.7≤SOC t ≤0.3, the unit operating cost of discharge is r2; if 0.3<SOC t≤0, the unit operating cost of discharge is r3. Where r1<r2<r3 and r1, r2, r3 are generally taken as 0.05, 0.1, and 0.25.

[0103] (2) When BESS is charging, if 0.7<SOC t ≤1, the unit operating cost of charging is r3; if 0.7≤SOC t ≤0.3, the unit operating cost of charging is r2; if 0.3<SOC t ≤0, the unit operating cost of charging is r1.

[0104] As can be seen from this, the BESS operating cost function is a piecewise function. This means that each calculation of the BESS operating cost during the real-time scheduling optimization process requires the execution of a judgment statement to determine the SOC region, which reduces the speed of real-time scheduling optimization. Therefore, to improve the calculation speed of the BESS real-time scheduling operating cost, the following linear model is established:

[0105]

[0106]

[0107] Where, is the real-time dispatching operation cost of BESS. is the average unit operating cost of BESS. SOC t is the state of charge of the BESS at time t. E is the maximum energy storage capacity of the BESS. Δt is the time interval between each dispatch. t RBESS It is the charging / discharging power scheduled by BESS in real time.

[0108] Step 3.2: Linear model of MT

[0109] The operating costs of a mobile operator (MT) consist of maintenance and fuel costs. Fuel costs are related to the unit price of fuel, the lower heating value of natural gas, operating efficiency, and power output. MT operating efficiency is a cubic function of power output. Calculating the operating costs of an MT requires first calculating its operating efficiency. To improve the calculation speed of MT real-time scheduling operating costs, the following linear model is established:

[0110]

[0111] Where, K is the real-time scheduling operation cost of MT. MT and λ are the maintenance coefficient of MT and the unit price of fuel respectively. t RMT Power output scheduled by MT in real time.

[0112] 3.3 Linear Model of CR

[0113] The CR in an island MG is used only for real-time scheduling. The CR is used to absorb excess power in the island MG, so that the island MG system achieves real-time power balance. Converting the excess power absorbed by the CR in the island MG into power loss cost, the real-time scheduling operating cost of the CR can be defined as the following linear model:

[0114]

[0115] Where, is the real-time dispatching cost of CR. τ is the unit cost of power loss of isolated MG. t CR is the amount of excess power consumed by CR in real-time scheduling.

[0116] Step 3.4: Real-time scheduling linear objective function of isolated MG

[0117] The real-time scheduling of the isolated MG is only related to the BESS, MT, and CR. Therefore, the real-time scheduling operation cost of the isolated MG is composed of the real-time scheduling operation costs of the BESS, MT, and CR. The following objective function is established:

[0118]

[0119] Where, The operating cost of real-time dispatch of isolated island MG energy. and They are all linear models, so Equation (5) is called the linear objective function of real-time scheduling of island MG.

[0120] Step 3.5: Linear constraints for real-time scheduling of isolated MGs

[0121] In addition to satisfying their own constraints during real-time scheduling optimization, BESS, MT, and CR also need to satisfy real-time power balance constraints. Therefore, the real-time scheduling linear model of the island MG needs to satisfy the following constraints:

[0122] (1) BESS charging and discharging power constraints and SOC constraints

[0123]

[0124] P t RBESS =P t aBESS +P t BESS (7)

[0125]

[0126] Where, and are the maximum charging power and maximum discharging power of BESS respectively, where P t RBESS A positive value indicates discharge. t aBESS and P t BESS The BESS dynamically adjusts power in real-time scheduling and the charge / discharge power plan in day-ahead scheduling. initial is the initial SOC value of BESS.

[0127] (2) MT power output constraints

[0128]

[0129] P t RMT =P t aMT +P t MT (10)

[0130] Where, is the maximum power output of MT. t aMT and P t MT MT dynamically adjusts power in real-time scheduling and plans power output in day-ahead scheduling.

[0131] (3) Quantity constraints on excess power absorbed by CR

[0132]

[0133] Where, It is the maximum amount of excess power that can be absorbed by CR.

[0134] (4) Real-time power balance constraints for real-time scheduling optimization of isolated MGs

[0135]

[0136]

[0137]

[0138] P t PLoad =P t Load -P t un (15)

[0139] P t PRES =P t RES -P t abRES (16)

[0140] Where, and are the deviation values ​​between the actual Load and RES in real-time scheduling and the Load and RES output plans in day-ahead scheduling optimization. t ALoad and P t PLoad They are the actual load in real-time scheduling and the load output plan in day-ahead scheduling optimization. t ARES and P t PRES They are the actual RES in real-time scheduling and the RES output plan in day-ahead scheduling optimization. t Load and P t un are the number of predicted load and unloaded load in the day-ahead scheduling optimization. t RES and P t abRES are the number of predicted RES and abandoned RES in day-ahead scheduling optimization, respectively.

[0141] In summary, the linear model for real-time scheduling of island MGs is linear, simple, and computationally inefficient. Therefore, the linear model for real-time scheduling of island MGs established in this invention is beneficial for improving the speed and ability of island MGs to handle the uncertainties of load and resource resources in real-time energy scheduling, and after optimization, a unique high-quality solution can be obtained.

[0142] Figure 4 is a simulation diagram of the real-time dispatch of isolated island MG energy. Figure 4a This is the BESS and MT output plan result diagram for the day-ahead dispatch of the isolated island MG energy; Figure 4b This is the result diagram of the unloading and abandonment of RES plan by the island MG Energy day-ahead dispatch; Figure 4c The actual output power of BESS and MT and the actual power consumption of CR for real-time dispatch of isolated island MG energy are shown in the figure. Figure 4d Optimize the time distribution diagram for the 24-hour energy real-time dispatch of the isolated island MG. Figure 4a and Figure 4cIt can be seen that the actual output power of the BESS and MT scheduled in real time is different from the output power plan of the BESS and MT scheduled in the day-ahead schedule, and the CR needs to be started to absorb the excess power in the real-time schedule. Therefore, the linear optimization method of the real-time scheduling of the energy of the isolated island MG of the present invention improves the real-time processing load and RES capability of the isolated island MG, thereby improving the reliability of the stable operation of the isolated island MG. Figure 4d As can be seen, at time t = 15, the island MG real-time energy scheduling optimization time is 1.28 seconds; at other times, the island MG real-time energy scheduling optimization time is completed within 1 second. Therefore, the linear optimization method for island MG real-time energy scheduling of the present invention improves the speed of island MG real-time processing of load and RES.

Claims

1. A linear optimization method for real-time scheduling of energy in an isolated microgrid, characterized in that: The following steps are involved: Step 1, design an island microgrid structure based on a multi-agent system: its structure includes renewable energy, batteries, micro gas turbines, loads and absorption resistors; its renewable energy, batteries, micro gas turbines, loads and absorption resistors are interconnected through power transmission lines; renewable energy, batteries, micro gas turbines, loads and absorption resistors are respectively connected to corresponding agents, wherein the agents of renewable energy, batteries, micro gas turbines, loads and absorption resistors are respectively connected to the microgrid agent; Step 2, designing a real-time dispatch optimization method for the isolated microgrid: including collecting the deviation value between the day-ahead load and the real-time load, and the deviation value between the day-ahead renewable energy and the real-time renewable energy; Then the load agent and the renewable energy agent send the obtained error values ​​to the microgrid agent respectively; after receiving the error values ​​from the load agent and the renewable energy agent, the microgrid agent calculates the total error value, namely: [actual load - day-ahead load] - [actual renewable energy - day-ahead renewable energy], and sends the total error value to the real-time scheduling linear model of the isolated microgrid; sends the day-ahead battery charge / discharge power plan and the day-ahead micro gas turbine power output plan to the real-time scheduling linear model of the isolated microgrid energy; performs energy real-time scheduling optimization, obtains the actual output plan of each unit and feeds it back to each unit; Step 3: Establish a linear model for real-time dispatch of the isolated island microgrid, including a linear model of the battery, a linear model of the micro gas turbine, a linear model of the absorption resistor, and a linear objective function for real-time dispatch of energy of the isolated island microgrid; Step 3.

1. Linear model of battery The operating cost of the battery is related to the charge / discharge power and the state of charge. The unit operating cost of each working area of ​​the battery can be set separately for the discharge and charge processes: (1) When the battery is discharging, if 0.7<SOC t ≤1, the unit operating cost of discharge is r1; if 0.7≤SOC t ≤0.3, the unit operating cost of discharge is r2; if 0.3<SOC t ≤0, the unit operating cost of discharge is r3; where r1<r2<r3 and r1, r2, r3 are generally taken as 0.05, 0.1, 0.25; (2) When the battery is charging, if 0.7<SOC t ≤1, the unit operating cost of charging is r3; if 0.7≤SOC t ≤0.3, the unit operating cost of charging is r2; if 0.3<SOC t ≤0, the unit operating cost of charging is r1; It can be seen that the operating cost function of the battery is a piecewise function. In order to improve the calculation speed of the real-time scheduling operating cost of the battery, the following linear model is established: In the formula, is the real-time dispatching operation cost of the battery, is the average unit operating cost of the battery, SOC t is the state of charge of the battery at time t, E is the maximum energy storage capacity of the battery, Δt is the interval between each dispatch, and P t RBESS Real-time scheduling of charging / discharging power for batteries; Step 3.2: Linear model of microturbine The operating cost of a micro gas turbine includes maintenance cost and fuel cost. Among them, the fuel cost is related to the unit price of fuel, the low calorific value of natural gas, operating efficiency and power output. The operating efficiency of a micro gas turbine is a cubic function of power output. First, the operating efficiency of a micro gas turbine must be calculated before the operating cost of the micro gas turbine can be calculated. In order to improve the calculation speed of the real-time scheduling operating cost of a micro gas turbine, the following linear model is established: In the formula, is the real-time dispatching operation cost of the micro gas turbine, K MT and λ are the maintenance factor and fuel unit price of the micro gas turbine, respectively. t RMT Real-time dispatch of power output for micro gas turbines; Step 3.3: Linear model of the absorber resistor The excess power absorbed by the absorption resistor in the isolated island microgrid is converted into power loss cost, and the real-time dispatching operation cost of the absorption resistor can be defined as the following linear model: In the formula, is the real-time dispatching operation cost of the absorption resistor, τ is the unit cost of power loss of the isolated microgrid, P t CR To absorb the amount of excess power consumed by the resistor in real-time scheduling; Step 3.4: Real-time dispatch linear objective function of isolated microgrid The real-time dispatching and operating costs of the isolated microgrid, including the real-time dispatching and operating costs of batteries, micro gas turbines and absorption resistors, establish the following objective function: In the formula, The operating cost of real-time dispatch of energy for isolated microgrids is and They are all linear models, so formula (5) is called the real-time scheduling linear objective function of the island microgrid.

2. According to claim 1, a linear optimization method for real-time scheduling of energy in an isolated island microgrid is characterized in that: The design of step 1 is based on the island microgrid structure of the multi-agent system. The process includes: In the island microgrid structure based on the multi-agent system, the agent connected to the renewable energy agent is used to collect the power output information of the renewable energy in real time; the agent connected to the battery is used to collect the state of charge of the battery, the upper and lower limits of the charge and discharge power, and the energy storage capacity information in real time, and send the charge / discharge power instruction to the battery; the agent connected to the micro gas turbine is used to collect the power output information and power output constraints of the micro gas turbine in real time, and send the power output instruction to the micro gas turbine; the agent connected to the load is used to collect the value of the load at each moment in real time; the agent connected to the absorption resistor is used to collect the number of absorption resistors and the constraint information of the number of excess power to be absorbed in real time, and send the amount of excess power to be consumed to the absorption resistor; the microgrid agent is responsible for real-time energy scheduling optimization; the agent connected to the renewable energy, battery, micro gas turbine, load and absorption resistor needs to send all the collected data information to the microgrid agent; the energy scheduling instruction optimized by the microgrid agent is sent to the agent connected to the battery, micro gas turbine and absorption resistor.

3. According to claim 1, a linear optimization method for real-time scheduling of energy in an isolated island microgrid is characterized in that: The specific process of the step 2 to design the real-time dispatch optimization method of the isolated island microgrid is as follows: First, the load agent collects the actual load and the day-ahead load and calculates their error value; the renewable energy agent collects the actual renewable energy and the day-ahead renewable energy and calculates their error value; At the same time, the battery agent and the micro gas turbine agent respectively send the battery charging / discharging power plan and the micro gas turbine power output plan in the day-ahead scheduling optimization to the real-time scheduling linear model of the isolated microgrid; then, the microgrid agent uses the linear programming solver to perform energy real-time scheduling optimization on the real-time scheduling linear model to obtain the actual output plan of each unit; the microgrid agent sends the actual output plan of each unit to the battery agent, the micro gas turbine agent and the absorption resistor agent respectively; the battery agent, the micro gas turbine agent and the absorption resistor agent send the corresponding actual output plan instructions to the battery, the micro gas turbine and the absorption resistor; finally, the battery, the micro gas turbine and the absorption resistor work according to the actual output plan.

4. According to claim 1, a linear optimization method for real-time scheduling of energy in an isolated island microgrid is characterized in that: In step 3: Step 3.5: Real-time dispatch linear constraints of isolated microgrid In real-time scheduling optimization, batteries, micro gas turbines and absorption resistors need to meet not only their own constraints but also real-time power balance constraints. Therefore, the real-time scheduling linear model of the island microgrid needs to meet the following constraints: (1) Battery charging and discharging power constraints and charge state constraints In the formula, and are the maximum charging power and maximum discharging power of the battery, respectively, where P t RBESS A positive value indicates discharge, P t aBESS and P t BESS The battery dynamically adjusts power in real-time scheduling and the charge / discharge power plan in day-ahead scheduling. SOC initial is the initial value of the battery's state of charge; (2) Power output constraints of micro gas turbines P t RMT =P t aMT +P t MT (10) In the formula, is the maximum power output of the micro gas turbine, P t aMT and P t MT It is the real-time dynamic adjustment of power of the micro gas turbine in real-time scheduling and the power output plan in day-ahead scheduling; (3) Constraints on the number of resistors that can absorb excess power In the formula, is the maximum amount of excess power that can be absorbed by the absorption resistor; (4) Real-time power balance constraints for real-time dispatch optimization of isolated microgrids P t PLoad =P t Load -P t un (15) P t PRES =P t RES -P t abRES (16) In the formula, and are the deviations between the actual load and renewable energy in real-time scheduling and the load and renewable energy output plan in day-ahead scheduling optimization, respectively. t ALoad and P t PLoad are the actual load in real-time dispatch and the load output plan in day-ahead dispatch optimization, P t ARES and P t PRES are the actual renewable energy in real-time scheduling and the renewable energy output plan in day-ahead scheduling optimization, P t Load and P t un are the number of forecast load and unloaded load in the day-ahead dispatch optimization, P t RES and P t abRES They are the number of predicted renewable energy and abandoned renewable energy in day-ahead dispatch optimization.

5. According to claim 1, a linear optimization method for real-time scheduling of energy in an isolated island microgrid is characterized in that: The algorithm of the linear programming solver in step 2 comprises the following steps: 1) Input: (1) Battery parameter input: maximum energy storage capacity E, each scheduling interval Δt, maximum charge / discharge power Unit operating cost r1, r2, r3, initial state of charge value SOC initial ; (2) Micro gas turbine parameter input: maintenance factor K MT , fuel unit price λ, maximum power output (3) Absorption resistance parameter input: unit cost of power loss of the isolated microgrid τ, maximum amount of excess power to be absorbed (4) Other parameter input: deviation value and Battery charging / discharging power plan P in day-ahead dispatch optimization t BESS and micro gas turbine output plan P t MT ; 2) Real variable setting: The battery dynamically adjusts the power P in real time t aBESS , Micro gas turbine real-time dynamic adjustment of power P t aMT 、The amount of excess power absorbed by the absorption resistor P t CR ; 3) Constraint setting: Formula (6): Formula (8): 0≤SOC t ≤1, formula (9): Formula (11): Formula (12): 4) Objective function setting: 5) Optimization solution: 6) Output: The battery dynamically adjusts the optimal power value in real time Real-time dynamic adjustment of the power optimum value of micro gas turbine The optimal value of the number of resistors that can absorb excess power

Images