A power distribution automation system collaborative planning method and device for distributed power supply access

Abstract

The application discloses a distribution automation system collaborative planning method and device for distributed power supply access. A DG-AVVC-AFM collaborative planning optimization model is constructed, a total cost minimum is taken as an objective function, and the total cost at least includes DG investment cost, automation device investment cost, operation power purchase cost, network loss cost and reliability cost based on un-supplied power EENS and VOLL; linearization processing is carried out on constraints including nonlinear power flow and capacity circular inequality, a mixed integer nonlinear programming model is converted into a mixed integer linear programming MILP model and is solved, and a DG location and capacity scheme, an AVVC location and capacity scheme and an AFM switch / interconnection line configuration scheme are output. The application can reduce operation cost and network loss and improve power supply reliability under the condition of guaranteeing radial structure and safe operation constraints, and has the advantages of high solving efficiency, global optimal or approximate global optimal solution.

Description

Technical Field

[0001] This invention belongs to the field of distribution network planning and optimization calculation technology, and relates to the collaborative planning of distribution automation system (DAS) under distributed generation (DG) access conditions. Specifically, it relates to the joint location and capacity determination of DG, automatic voltage and reactive power control (AVVC) device and automatic fault management (AFM) device, tie line / switch configuration, and the method and apparatus for linearizing the collaborative planning model containing nonlinear power flow and capacity constraints and solving it using mixed integer linear programming (MILP). Background Technology

[0002] Distribution automation systems can enhance the state awareness, fault isolation, and self-healing capabilities of distribution networks, and improve power supply reliability. Distributed generation (DG) integration can improve node voltage distribution, reduce network losses, and enhance power supply reliability. However, with large-scale DG integration, traditional planning methods focusing on a single device or a single objective (such as economic efficiency alone) struggle to balance operating costs, network losses, reliability, and voltage quality. Furthermore, collaborative planning models typically exhibit mixed-integer nonlinear programming (MINLP), which suffers from computational time consumption, parameter sensitivity, and susceptibility to local optima when solved using heuristics or evolutionary algorithms, and it is difficult to provide verifiable optimality definitions. Therefore, a planning method is needed that can simultaneously consider the collaborative configuration of DG and distribution automation devices, while offering higher solution efficiency and guaranteed solution quality. Summary of the Invention

[0003] The purpose of this invention is to provide a collaborative planning method and apparatus for distribution automation systems with distributed power source access. (1) Implement joint planning of DG, AVVC, and AFM, taking into account economy, loss and reliability; (2) Linearize the MINLP model containing nonlinear power flow and capacity constraints into the MILP model to improve the solution efficiency and increase the possibility of obtaining the global optimum or near-global optimum solution; (3) Under the conditions of meeting the investment budget, operating limits, voltage constraints and radial structure, output a feasible collaborative planning scheme. Attached Figure Description

[0004] Figure 1 Schematic diagram of power distribution network topology; Figure 2 : A schematic diagram of the collaborative planning results output by this invention (DG, AVVC installation location and capacity, AFM tie line / switch configuration). Figure 3 Comparison of power curves under different scenarios; Figure 4 Comparison of node voltage distribution under peak and non-peak conditions; Figure 5 : Reliability cost versus EENS variation under different VOLL values. Detailed Implementation

[0005] Step S1: Data Acquisition and Candidate Set Construction Inputs: Distribution network topology (node ​​set, line set, candidate tie line set / switchable line set); line parameters (conductance / susceptance, etc.), substation capacity limit; load demand for each node (given or generated by time t, scenario s); allowable voltage range for nodes (minimum / maximum voltage amplitude) and slack node settings; candidate installation location set: DG candidate nodes, AVVC candidate nodes, AFM candidate line / switch locations; cost and economic parameters: time-of-use tariff, DG unit operating cost; scenario set and its probability of occurrence; reliability parameter: loss-of-load value (VOLL). Among the above parameters, time-of-use tariff, budget, and VOLL can be set according to the project / example.

[0006] Output: Network data, candidate set, and parameter set used for optimization modeling.

[0007] Step S2: Establish the mapping between collaborative planning objects and functions: Input: The candidate set and parameters output from step S1.

[0008] Handling: Clearly define the planning objects: AVVC system: regulates voltage and reactive power through capacitor banks, static var compensators, and voltage / reactive power control equipment; AFM system: includes protective switches, tie lines, and switches, used for fault isolation / network reconfiguration and reliability improvement; DG: can inject power into the distribution network to improve operation and reliability when load demand is met.

[0009] Output: The correspondence between the three types of equipment / measures (DG, AVVC, AFM) and system operating indicators (voltage, capacity, reliability).

[0010] Step S3: Construct a system of decision variables (planning variables and operational variables): Input: Steps S1 and S2 output.

[0011] Processing: Definitions must include at least: Investment / Construction Decisions (binary variables): DG construction variable (construction = 1 / no construction = 0); AVVC construction variable (construction = 1 / no construction = 0); AFM construction variable (construction = 1 / no construction = 0). Capacity / Configuration Decisions (continuous variables): DG capacity, AVVC capacity, etc.; Operating status variables (by time t, scenario s): Substation output active / reactive power; DG output active / reactive power; line active / reactive power flow; node voltage amplitude and phase angle; binary variables related to AFM operating status regarding line opening / closing (open = 1 / closed = 0).

[0012] Output: A set of variables that can be used to express the objective function and constraints.

[0013] Step S4: Construct the objective function (minimize total cost): Input: Set of variables and cost parameters.

[0014] Processing: Establish a "minimize total cost" objective function, which includes at least: DG investment cost (related to DG construction variables and capacity); automation device investment cost (AVVC and AFM investment); distribution system operation cost (related to time-of-use pricing, DG unit operation cost, power purchase at slack nodes, DG output power, and scenario probability); energy loss cost (related to network loss / load); reliability cost: constructed based on load shedding EENS and offload value VOLL.

[0015] Output: The objective function expression for collaborative programming optimization.

[0016] Specifically, this invention targets a Distribution System (DAS) that includes Automatic Voltage and VAR Control (AVVC) and Automatic Fault Management (AFM) systems. The AVVC system provides the use of capacitor banks, static var compensators, and voltage and reactive power control equipment to regulate voltage and reactive power in the distribution network; the AFM system includes protective switches, tie lines, and switches. This invention proposes a distribution network collaborative planning model that includes distributed generation, the AVVC system, and the AFM system. The investment cost of distributed generation is considered. Investment cost of automatic protection devices for power distribution networks Operating costs of power distribution systems Energy loss cost And reliability costs based on load shedding and underload values. The objective function for minimizing the minimum value can be expressed as: (1); Among them, the investment cost of distributed power sources for: (2); In the formula: For system nodes n A set; For DG's investment costs; The variable is 0-1, where 1 indicates the investment in DG. A value of 0 indicates that no DG will be invested in.

[0017] Investment cost of automatic protection devices for power distribution networks for: (3); In the formula: Investment cost of AVVC equipment; The variable is 0-1, with a value of 1 indicating the installation of an AVVC device. A value of 0 indicates that no AVVC device will be installed; Investment cost of AFM unit; The variable is 0-1, where 1 indicates the deployment of an AFM device. A value of 0 indicates that no AFM unit will be built.

[0018] Operating costs of power distribution systems for: (4); In the formula: These are collections of scenes and times, respectively. For electricity price; DG unit operating cost; For time t node n Scene s The active power output of the distributed power source; For time t Active power at the relaxation node; For the scene s The probability of occurrence.

[0019] Energy loss cost for: (5); In the formula: Let t be the load demand at node n in scenario s.

[0020] Reliability costs based on Expected Energy Not Supplied (EENS) and Value Of Lost Load (VOLL) for: (6); In the formula: VOLL is the offload value; This refers to the active power not provided at time node t in scenario s.

[0021] Step S5: Construct a set of constraints: Input: Objective function and variables.

[0022] Processing: At least the following constraints should be established (each type of constraint can be expanded according to scenario s and time t): Investment budget constraint: DG, AVVC, and AFM investments should not exceed the corresponding maximum investment cost / budget. Power balance constraint (node ​​active / reactive power): Active / reactive power balance should be satisfied at each node, at each time, and in each scenario, and an unsupplied active power variable should be introduced to represent load shedding. Line injection power constraint (power flow equation): Line active / reactive power is related to conductance susceptance, node voltage amplitude, and phase angle, and line opening and closing are controlled by binary variables of line opening related to AFM operating status. Maximum operating limit constraint: Maximum load / capacity limits of lines and substations, and coupling constraints between the maximum capacity of DG that can be installed at nodes and DG deployment variables. Node voltage constraint: Upper and lower limits of node voltage amplitude, and relaxed node voltage reference constraints. AVVC capacity constraint: Upper and lower limits of AVVC capacity and coupling with AVVC deployment variables. AFM protection switch and radial structure constraint: Including the constraint of the maximum number of allowed switching operations on the line and the constraint of ensuring the radial structure of the distribution network (related to the number of nodes, closed branches, etc.). Reliability constraints: The power not supplied must not exceed the load requirement, etc.

[0023] Output: A complete set of collaborative planning constraints.

[0024] Specifically: 1) Investment cost constraints: This invention considers the maximum investable budget of the investment entity during the planning process. The investment cost constraints for distributed power sources, AVVC devices with automatic voltage and reactive power control, and AFM automatic fault management devices are as follows: (7); (8); (9); In the formula: This represents the maximum investment cost of distributed power sources. This represents the maximum investment cost for an AVVC device. This represents the maximum investment cost of an AFM unit.

[0025] 2) Power balance constraints: (10); (11); In the formula: For time t node n Scene s The active power not provided below; , They are time t node n Scenes The active and reactive power output of the substation; , They are time t node n Scene s The active and reactive power output of the distributed power source; , They are time t Scene s Next node n To node j Active power flow and reactive power flow; For nodes n , j The correlation matrix; , They are time t node n Scene s The active and reactive load demands.

[0026] 3) Constraints on active and reactive power injected into the line: (12); (13); In the formula: , They are nodes n With nodes j Conductivity and susceptance between lines; , They are time t Scene s Next node n With nodes j The voltage; , They are time t Scene s Next node n With nodes j The phase angle; This is a binary variable related to the operating status of the AFM device. A value of 1 indicates the on state, and a value of 0 indicates the off state.

[0027] 4) Maximum operating constraints of the power distribution system: (14); (15); (16); In the formula: , The lines are respectively n,jand the maximum load of the substation; For nodes n The maximum capacity of the distributed power source that can be installed at the location; For the binary variable related to the DG investment status, a value of 1 indicates the construction of distributed power generation, and the node is... n This is a suitable location to install DG. A value of 0 indicates that no distributed power generation will be invested in.

[0028] 5) Node voltage constraints: (17); (18); In the formula: , These represent the minimum and maximum voltage values ​​allowed for node n, respectively; ref indicates a relaxed node.

[0029] 6) AVVC planning capacity constraints: (19); (20); In the formula: For nodes n The maximum capacity of AVVC that can be built; For AVVC investment status, a binary variable is defined as 1, indicating investment in AVVC. (Node) n This is the appropriate location to install AVVC. A value of 0 indicates that no AVVC will be built.

[0030] 7) AFM protection switch constraints and radial structural constraints: (twenty one); (twenty two); In the formula: This represents the maximum number of switching operations allowed on line nj. This represents the total number of nodes in the distribution network.

[0031] 8) Reliability constraints: (twenty three); In the formula: For active power not provided in scenario s at time node t; power; Let t be the load demand at node n in scenario s.

[0032] Step S6: Form MINLP and identify nonlinear / nonconvex constraints: Input: Objective function and set of constraints.

[0033] Solution: Since the model contains nonlinear constraints (power flow related constraints, capacity constraints) and nonconvex equations coupled with binary variables, it belongs to MINLP as a whole; traditional numerical methods or evolutionary algorithms may lead to local optima and slow computation speed.

[0034] Output: The MINLP model to be linearized and a list of its nonlinear / nonconvex parts.

[0035] Specifically: Because the proposed planning model includes nonlinear constraints (12)-(16), nonconvex equations (10)-(11), and binary variables , , and Therefore, this model is a MINLP model. Although it can be solved using traditional numerical methods such as the Newton-Raphson method or evolutionary algorithms, this will result in the optimal solution in the non-convex equation system being a local optimum and the computation speed being slow. To solve the above-mentioned problem, this invention will transform the proposed MINLP model into a Mixed Integer Linear Program (MILP) model through linearization to achieve higher solution accuracy and efficiency.

[0036] Step S7: Linearization and conversion to MILP: Input: MINLP model and a list of nonlinear / nonconvex parts.

[0037] Processing: Perform at least two types of linearization: (1) Linearization of power flow triangle terms: When the voltage angle difference between adjacent busbars connected by the line is less than 0.105 rad, the trigonometric function terms in the power flow expression are approximated (for example, the cosine / sine terms are approximated by 1 and the angle difference), and parameters such as constant M and line slope are introduced to achieve linear coupling with the line interruption state variables, thereby transforming the relevant nonlinear power flow and voltage constraints into a linear form.

[0038] Specifically, when the voltage angle difference between two adjacent busbars connected by a line is less than 0.105 rad, the values ​​in equations (12)-(13) are... and You can use 1 and To replace, , and They can be represented as follows: and Therefore, equations (12)-(13) and (17) can be transformed into: (twenty four); (25); (26); In the formula: For the set of lines; M is a constant; m is the slope of the line; This refers to the voltage deviation.

[0039] (2) Piecewise linearization of the capacity “circular inequality”: For circular inequalities representing maximum apparent power limits (such as line / substation capacity constraints), the circumference is divided into k equal parts, and multiple linear piecewise constraints with different angles are constructed. The polygon approximates the circular boundary, reducing linearization errors and obtaining a set of linear inequalities.

[0040] Equations (14)-(16) are circular inequalities, representing the intervals of change between active and reactive power within a circle with a radius equal to the maximum apparent power. The error value will be high with a small number of piecewise linear segments. The number of piecewise linear segments at different angles to the horizontal axis can be increased to reduce linearization error. Therefore, this invention divides the circumference into k equal segments. Integrating the calculated equation of the straight line into a circle with a radius less than or equal to S, the linear approximation is: (27); (28); (29); After linearization, the original MINLP model is rewritten as a MILP model.

[0041] Output: MILP programming model that can be directly solved by the MILP solver.

[0042] The linearized MILP programming model can be rewritten as: (30); Step S8: MILP solution and collaborative planning scheme output: Input: MILP model.

[0043] Processing: The MILP solver is invoked to obtain the optimal / optimal feasible solution. Output: DG installation nodes and capacity (site selection and capacity); AVVC installation nodes and capacity; AFM configuration results (tie line location, switch configuration and on / off status / operable strategy); and further calculation of comparative indicators (investment cost, operating cost, energy loss, reliability cost, maximum voltage deviation, EENS, etc.).

[0044] Output: A complete comparison of collaborative planning results and performance indicators; for example, it can show significant reductions in operating costs, losses and reliability indicators, as well as reductions in maximum voltage deviation and EENS.

[0045] Example: To verify the effectiveness of the method proposed in this invention, simulation analysis was performed using the IEEE 69-node system, whose topology is as follows: Figure 1 As shown. The maximum capacity of the AVVC device connected to the distribution network is 200 kvar, with a unit investment cost of 650,000 yuan; the maximum capacity of the distributed generation connected to the distribution network is 350 kVA, with unit investment costs and operating costs of 3.68 million yuan and 140 yuan / MWh, respectively. The investment budgets for installing DG, AVVC, and AFM in the above distribution network system are 8 million yuan, 2 million yuan, and 1 million yuan, respectively; the electricity price is 112 yuan / MWh during 1:00-7:00, 168 yuan / MWh during 8:00-17:00, and 215 yuan / MWh during 23:00-24:00 and 18:00-22:00; VOLL is 650 yuan / MWh. Two scenarios are set up for comparative analysis, where scenario 1 is the distribution network planning method without considering distributed generation; scenario 2 is the method of this invention, namely the distribution network collaborative planning method considering distributed generation.

[0046] The planning results for scenario 2 are as follows Figure 2 As shown. From Figure 2 As can be seen from the planning method proposed by the present invention, four tie lines are installed in the power distribution system to improve the reliability of the system, located at nodes 11-40, 15-67, 24-61 and 35-56; AVVC systems are installed at nodes 24, 35, 45 and 65 to improve the voltage distribution of the system; and six distributed power sources are installed at nodes 18, 48, 55, 58, 60 and 64.

[0047] The economic comparison results of Scenario 1 and Scenario 2 are shown in Table 1. It can be seen that, compared with Scenario 1, the total investment cost of Scenario 2 is 5.8341 million yuan, but the operating cost, energy loss cost and reliability cost after investment are significantly reduced. The proposed planning results reduce the total operating cost by 33.4% compared with the previous one, which has good economic efficiency.

[0048] Table 1. Comparison of Planning Economics: Tab. 1 Comparison of Planning Economics: ; Table 2 shows the comparison results of the system's maximum voltage deviation, energy loss, and EENS. It can be seen that, compared with scenario 1, scenario 2 has significantly reduced maximum voltage deviation, energy loss, and EENS, indicating that the planning results have better operating characteristics and reliability.

[0049] Table 2 Comparison of System Performance: ; Comparison of operating power between Scenario 1 and Scenario 2 Figure 3 As shown in the diagram, during the period from 18:00 to 22:00, Scenario 1 experienced a power overload, with the operating power exceeding the maximum allowable value. Compared to Scenario 1, the apparent power in Scenario 2 generally decreases, placing it within a safer operating range. This is because of the presence of distributed power sources in the distribution network protection device, which can inject power into the distribution network when needed to meet load demands, thereby improving system reliability and promoting stable system operation.

[0050] The node voltage distribution of the system under peak and non-peak load conditions, for example Figure 4 As shown, the non-peak load is taken as 40% of the peak load. It can be seen that installing distributed generation devices at the obtained location can improve the voltage amplitude and keep it within an acceptable range. Compared with scenario 1, the voltage distribution under the method proposed in this invention is flatter, and a reasonable planning scheme can effectively improve the voltage distribution.

[0051] The reliability cost and EENS variation with VOLL of the loss of load in scenario 2 are as follows: Figure 5 As shown, EENS decreases with increasing VOLL, reaching its lowest value when VOLL is 1000 yuan / MWh. Reliability cost initially increases and then decreases with increasing VOLL, steadily increasing from zero until reaching a peak at VOLL of 400 yuan / MWh, then decreasing until reaching zero expected cost at VOLL of 1000 yuan / MWh. This demonstrates that the planning model based on this invention can effectively improve system reliability and operational indicators.

[0052] Beneficial effects: Compared with the prior art, the present invention has at least the following beneficial effects: (1) It realizes the collaborative planning of DG and distribution automation device AVVC / AFM, comprehensively considers operating costs, network loss costs and reliability costs, and outputs a comprehensive optimal solution that is more in line with the background of high DG penetration; (2) By linearizing the nonlinear / nonconvex constraints of power flow and capacity, MINLP is transformed into MILP, which improves the solution efficiency, reduces the risk of getting trapped in local optima, and is more conducive to obtaining a solution with verifiable optimality; (3) Under the condition of satisfying the safety constraints such as voltage constraints, capacity constraints and radial structure, it can significantly improve voltage distribution and reduce EENS, thereby improving power supply reliability. The present invention can be deployed in distribution network planning software, distribution automation master station / planning platform or power grid enterprise planning calculation system, and is applicable to the planning and transformation decision of distribution automation system under the condition of high DG penetration access, and has good engineering feasibility and promotion value.

Claims

1. A collaborative planning method and apparatus for a distribution automation system with distributed power source access, characterized in that, Includes the following steps: S1. Obtain distribution network topology data, line parameters, substation capacity limit, node load demand data, node voltage allowable range, candidate installation location set, as well as cost parameters, investment budget parameters, time-of-use pricing information, scenario set and its occurrence probability, and offload value VOLL for distributed generation (DG), automatic voltage and reactive power control (AVVC), and automatic fault management (AFM). S2. Establish the mapping relationship between DG, AVVC, and AFM collaborative planning objects and functions; S3. Construct a decision variable system for the collaborative planning optimization model. The decision variable system includes at least binary variables of DG commissioning status, binary variables of AVVC commissioning status, binary variables of AFM commissioning status, as well as DG capacity variables, AVVC capacity variables, binary variables of line interruption status, and node voltage variables. S4. Construct a collaborative planning optimization model with the objective function of minimizing total cost, wherein the total cost includes at least the DG investment cost, the automation device investment cost, the power distribution system operation cost, the energy loss cost, and the reliability cost based on the unsupplied power EENS and VOLL. S5. Construct a set of constraints for the collaborative planning optimization model. The set of constraints includes at least investment budget constraints, node power balance constraints, line injection power constraints, maximum limit operation constraints of the distribution system, node voltage constraints, AVVC capacity constraints, AFM protection switch and radial structure constraints, and reliability constraints. S6. Identify the nonlinear / nonconvex constraints in the collaborative planning optimization model and form a mixed integer nonlinear programming (MINLP) model. S7. The MINLP model is linearized and transformed into a mixed integer linear programming (MILP) model. The linearization process includes at least power flow triangulation and capacity circular inequality piecewise linearization. S8. Call the MILP solver to solve the MILP model and output the DG addressing and sizing scheme, AVVC addressing and sizing scheme, and AFM switch / tie line configuration scheme.

2. The method according to claim 1, characterized in that: The operating cost of the distribution system is related to time-of-use pricing, unit operating cost of distributed generation (DG), active power purchased by slack nodes, active power output of DG, and the probability of scenario occurrence. The reliability cost is constructed using load shedding (EENS) and off-load value (VOLL), where EENS is accumulated from the active power not supplied under various time scales and scenarios. The investment budget constraints include at least the maximum investment cost constraints for DG, AVVC, and AFM. The node power balance constraints include active and reactive power balance constraints for each node under various scenarios and time periods, and introduce an active power variable to represent load shedding. In the line injection power constraints, the line active / reactive power is related to the line conductance, node voltage amplitude, and phase angle, and the line opening or closing state is controlled by a binary variable related to the AFM operating state. The maximum limit operating constraints of the distribution system include at least the maximum line load constraint, the maximum substation load constraint, and the maximum capacity constraint for DG installation at nodes, and the coupling of capacity constraints and deployment decisions is achieved through a binary variable of DG deployment status. The node voltage constraints include upper and lower limits for node voltage amplitude and relaxed node voltage reference constraints. The AVVC capacity constraints include the maximum deployable AVVC at the node, and the capacity constraints and deployment decisions are coupled through binary variables of the AVVC deployment status.

3. The method according to claim 1, characterized in that: The AFM protection switch and radial structure constraints include at least the maximum number of switching operations allowed on the line and the constraint of maintaining the radial structure of the distribution network.

4. The method according to claim 1, characterized in that: The power flow trigonometric linearization includes approximating the trigonometric function terms in the power flow expression when the voltage angle difference between adjacent buses is less than a preset threshold, and introducing a constant M and a line slope parameter to achieve linear coupling with the binary variables of the line interruption state, thereby transforming the nonlinear power flow constraint and voltage constraint into a linear form.

5. The method according to claim 11, characterized in that: The preset threshold is 0.105 rad.

6. The method according to claim 1, characterized in that: The segmented linearization of the capacity circular inequality includes dividing the circumference into k equal parts, constructing multiple linear segmented constraints at different angles to the horizontal axis, approximating the apparent power circular constraint boundary with a polygon, and forming a set of linear inequalities.

7. A collaborative planning device for a distribution automation system oriented towards distributed power source access, characterized in that, include: The data acquisition module is used to perform the data acquisition in step S1 of claim 1; The object mapping module is used to perform the collaborative planning object and function mapping in step S2 of claim 1; The variable construction module is used to perform the construction of the decision variable system in step S3 of claim 1; The objective function construction module is used to perform the objective function construction in step S4 of claim 1; A constraint construction module is used to perform the construction of the constraint set in step S5 of claim 1; A nonlinear identification module is used to perform the MINLP model formation and nonlinear / nonconvex constraint identification in step S6 of claim 1. The linearization module is used to perform the power flow triangulation and capacity circular inequality piecewise linearization in step S7 of claim 1, and to generate the MILP model. The solver and output module is used to execute step S8 in claim 1, call the MILP solver and output the DG addressing and sizing scheme, the AVVC addressing and sizing scheme and the AFM switch / tie line configuration scheme.

8. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method of any one of claims 1 to 7.

Images