Energy optimization scheduling method and system based on power distribution network-microgrid cooperation

Abstract

This invention discloses an energy optimization scheduling method based on distribution network-microgrid collaboration, including: acquiring data information of the target distribution network and microgrid; constructing a mathematical model and constraints for the distribution network-microgrid collaborative operation system; constructing an energy optimization scheduling objective function based on distribution network-microgrid collaboration; constructing a multi-agent structure of the distribution network-microgrid to form a multi-agent structure optimization scheduling model; designing a self-optimizing algorithm based on multi-agent differential evolution; and solving the multi-agent structure optimization scheduling model using the designed method to complete the energy optimization scheduling based on distribution network-microgrid collaboration. This invention also discloses a system for implementing the aforementioned energy optimization scheduling method based on distribution network-microgrid collaboration. This invention not only achieves energy optimization scheduling based on distribution network-microgrid collaboration but also offers higher reliability, better accuracy, and higher efficiency.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, specifically relating to an energy optimization scheduling method and system based on the coordinated operation of distribution network and microgrid. Background Technology

[0002] With economic and technological development and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and daily life, bringing endless convenience. Therefore, ensuring a stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] Currently, with the large-scale integration of distributed photovoltaic and wind power generation, microgrids and distribution networks have formed a hierarchical operation pattern of grid-connected coordination. Therefore, it is urgent to achieve coordinated and optimized scheduling of distribution networks and microgrids while meeting constraints such as AC power flow, voltage and line safety, energy storage state of charge, grid connection point and tie line capacity.

[0004] However, existing power system dispatching schemes all suffer from various problems: centralized optimization dispatching schemes are suitable for global dispatching, but under conditions of strong coupling and large-scale deployment, these schemes have high computational overhead, insufficient real-time performance, and are prone to conflicts between global objectives and local execution; distributed and decomposition optimization dispatching schemes rely heavily on convexity assumptions and precise parameters, so these schemes cannot guarantee the convergence of the scheme and the consistency of multiple stakeholders under non-convex equipment characteristics and communication disturbances; heuristic and evolutionary schemes can handle non-convex and mixed integer problems, but these schemes are mostly fixed parameter schemes, lacking safety awareness of voltage overruns and line overloads, and rarely take the consistency of grid connection point and tie line power as explicit constraints or costs, thus causing a disconnect between the global and local aspects and low convergence efficiency. Summary of the Invention

[0005] One of the objectives of this invention is to provide a highly reliable, accurate, and efficient energy optimization scheduling method based on the coordination of distribution network and microgrid.

[0006] The second objective of this invention is to provide a system for implementing the energy optimization scheduling method based on the coordinated operation of distribution network and microgrid.

[0007] The energy optimization scheduling method based on distribution network-microgrid coordination provided by this invention includes the following steps:

[0008] S1. Obtain data information from the target distribution network and microgrid;

[0009] S2. Based on the data obtained in step S1, construct the mathematical model and constraints of the distribution network-microgrid collaborative operation system;

[0010] S3. Based on the model and constraints established in step S2, construct an energy optimization scheduling objective function based on the coordination of distribution network and microgrid;

[0011] S4. Based on the objective function constructed in step S3, construct a multi-agent structure for the distribution network-microgrid to form an optimized scheduling model for the multi-agent structure;

[0012] S5. For the model obtained in step S4, design a self-optimization algorithm based on multi-agent differential evolution;

[0013] S6. Using the method designed in step S5, solve the multi-agent structure optimization scheduling model to complete the energy optimization scheduling based on the coordination of distribution network and microgrid.

[0014] Step S1, which involves acquiring data information about the target distribution network and microgrid, specifically includes the following steps:

[0015] Acquire data information from the target distribution network and microgrid;

[0016] The data information includes parameter data information of distributed power sources, operation data information of distributed power sources, operation data information of distribution networks and microgrids, and parameter data information of distribution networks and microgrids.

[0017] Step S2, which involves constructing a mathematical model and constraints for the distribution network-microgrid collaborative operation system based on the data information obtained in step S1, specifically includes the following steps:

[0018] The node power balance model is constructed using the following formula:

[0019] In the formula Let be the equivalent active load of node i at time t; For the set of distributed power sources connected to node i; Let g be the active power output of the distributed power source g at time t; Let be the energy storage discharge power of node i at time t; Let be the energy storage charging power of node i at time t; The active power exchange between node i and the upstream power grid or tie line at time t; Let i be the equivalent load at time t after considering flexible adjustment. Let be the equivalent reactive load of node i at time t; Let g be the reactive power output of the distributed power source g at time t. Let be the energy storage reactive power of node i at time t; The reactive power interaction between node i and the upstream power grid or tie line at time t; Let be the voltage amplitude at node i at time t; A set of nodes within the system; Let be the voltage amplitude at node j at time t; branch road The real part of the admittance; branch road The imaginary part of the admittance; Let be the voltage phase angle difference between node i and node j at time t;

[0020] The following formula is used as the active power output constraint for distributed generation:

[0021] In the formula Let g be the lower limit of the active power output of the distributed power source g at time t. Let g be the grid-connected state variable of the distributed power source g. If the distributed power source g is in the grid-connected state at time t, then... If the distributed power source g is in a non-grid-connected state at time t, then ; Let g be the active power output of the distributed power source g at time t; Let g be the upper limit of the active power output of the distributed power source g at time t; Let g be the reactive power output of the distributed power source g at time t. The rated capacity of the distributed power source g;

[0022] If the distributed power source g is a gas turbine, then the following additional ramp constraint is added:

[0023] In the formula The ramp-up limit for distributed power source g is reduced; The ramp-up limit for distributed power source g;

[0024] If the distributed power source g is distributed photovoltaic (PV), then the following formula is used as the output capacity constraint of the PV inverter:

[0025] In the formula Let be the photovoltaic active power available from distributed power source g at time t; The rated apparent capacity of the photovoltaic inverter corresponding to the distributed power source g; This is the voltage-reactive power droop factor; This is a reference value for the node voltage; This represents the lower limit of reactive power output of the photovoltaic inverter. This represents the upper limit of reactive power output of the photovoltaic inverter. It is a saturation function, and ;

[0026] For energy storage devices, the following formula is used as the energy storage system model:

[0027] In the formula Let be the maximum charging power of the energy storage device at node i at time t; Let be the charging state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a charging state at time t, then... If the energy storage device at node i is in a non-charging state at time t, then ; Let be the maximum discharge power of the energy storage device at node i at time t; Let be the discharge state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a discharge state at time t, then... If the energy storage device at node i is in a non-discharge state at time t, then ; Let t be the energy state of the energy storage device at time t; This refers to the charging power. This refers to the discharge power. To adjust the step size; This is the lower limit of the energy of the energy storage device; This represents the upper limit of the energy of the energy storage device.

[0028] For energy storage devices, the following formula is used as the reactive power constraint for energy storage:

[0029] In the formula Let be the net active power of the energy storage device at node i at time t, and ; Let be the rated apparent capacity of the energy storage device converter at node i;

[0030] For the flexible load at node i, the following formula is used as the flexible load model and constraint:

[0031] In the formula Let be the baseline load of node i at time t; Let be the adjustment amount at node i at time t; Let i be the load reduction limit for node i; Set the load increase limit for node i;

[0032] For transferable loads, the formula is used. As an energy conservation constraint within a given scheduling period T;

[0033] The following formula is used as the tie line and operational safety constraint:

[0034] In the formula Let be the active power of the tie line between microgrid m and the upper-level distribution network at time t; Let be the reactive power of the tie line between microgrid m and the upper-level distribution network at time t; The capacity of the tie line between the microgrid m and the upper-level distribution network; Let be the lower limit of the voltage at node i; This represents the upper limit of the voltage at node i; branch at time t The active power; branch at time t reactive power; branch road Maximum capacity.

[0035] Step S3, which involves constructing an energy optimization scheduling objective function based on the model and constraints established in step S2, specifically includes the following steps:

[0036] Within the scheduling period T, the following formula is used as the system's full-cycle decision vector. :

[0037] In the formula Let be the system decision vector at time t; Let t be the decision vector of the distribution network coordination layer; Let be the local decision vector of microgrid m at time t; For the system's microgrid collection; For a microgrid m, it is a collection of distributed power sources. A collection of energy storage devices within a microgrid m; The set of flexible load nodes within a microgrid m; This represents the power purchased by the distribution network and the upstream power grid at time t. This represents the electricity sales power of the distribution network and the upstream power grid at time t. The energy state of energy storage device s within microgrid m at time t;

[0038] The following formula is adopted as the objective function for energy optimization scheduling based on distribution network-microgrid coordination. :

[0039] In the formula The cost of purchasing and selling electricity from the upper-level power grid, and , The electricity purchase price between the distribution network and the upstream power grid at time t. The electricity price at time t represents the price at which the distribution network and the upstream power grid are sold. For the operating cost of distributed power sources, and , The first cost coefficient for secondary fuels The second cost coefficient for secondary fuels. The third cost coefficient for secondary fuels; The startup cost for distributed power sources. Let be the startup state variable of the distributed power source at time t. If the distributed power source is in the startup state at time t, then... If the distributed power source is in a non-startup state at time t, then... , The cost of downtime for distributed power sources. Let be the shutdown state variable of the distributed power source at time t. If the distributed power source is in a shutdown state at time t, then... If the distributed power source is in a non-shutdown state at time t, then , ; The costs associated with energy storage charging, discharging, and depreciation, and , Let be the charging power of the energy stored at time t, s. Let be the discharge power of the stored energy s at time t. The depreciation cost factor per unit charging capacity of energy storage s. The unit discharge capacity loss cost factor for energy storage s; For flexible load adjustment costs, and , This is the upward displacement relative to the reference load. This is the amount of reduction relative to the baseline load. Let i be the unit positive adjustment cost. The unit negative adjustment cost for node i; For network loss costs, and , For grid loss electricity price, For system network loss, , For the collection of distribution network branches, For microgrid branch collection, branch road The active power loss at time t, , branch road The equivalent resistance, branch road The meritorious trend, branch road The unproductive current; Weighting for voltage over-limit penalties; Let be the voltage over-limit penalty function, and , For positive part operators, , The square of the positive part operator; Weighting of branch circuit overload penalties; Let be the branch overload penalty function, and .

[0040] Step S4, which involves constructing a distribution network-microgrid multi-agent structure based on the objective function built in step S3 to form a multi-agent structure optimization scheduling model, specifically includes the following steps:

[0041] The agents are divided into groups, and the set of agents is represented as follows:

[0042] In the formula For distribution network coordination intelligent agents; For microgrid intelligent agents;

[0043] The following formula is used as Local cost function :

[0044] The following formula is used as Local cost function :

[0045] At this point, the global objective function Represented as ;

[0046] For a microgrid m, the active power injected at time t by the tie nodes, calculated according to the power flow equations, is... and injected reactive power Then the tie line power residual Represented as:

[0047] In the formula The active power residual of the tie line; For the reactive power residual of the tie line; if and only if At that time, the tie line power is the same as the injection power;

[0048] Constructing the tie-line power consistency penalty function for microgrid m , represented as ,in The square of the L2 norm;

[0049] The tie-line power consistency penalty term is distributed to the local objective function of each agent, and the augmented local objective function of each agent is set as follows:

[0050] In the formula For intelligent agents The augmented local objective function; This is the first non-negative weighting coefficient; For intelligent agents The augmented local objective function; This is the second non-negative weighting coefficient;

[0051] Finally, the multi-agent joint optimization model for micro-cooperative operation is expressed as:

[0052] In the formula This is the feasible region formed by the various constraints in step S2.

[0053] Step S5 describes designing a self-optimization algorithm based on multi-agent differential evolution for the model obtained in step S4, specifically including the following steps:

[0054] Algorithm initialization:

[0055] Set the evolution iteration number index k takes the value For any intelligent agent Set the population size as The local decision vector of the p-th individual in the k-th generation is represented as: ;

[0056] For any algebra The global reference solution is formed by combining the best local individuals of each agent. ;

[0057] Update adaptive parameters:

[0058] Constructing a normalized index based on global operational safety , is represented as:

[0059] In the formula This is the total global security penalty for the kth generation calculated according to step S3; This represents the total global security penalty corresponding to the 0th generation population. This is a set minimum value to prevent the denominator from being 0;

[0060] Define adaptive difference coefficient of variation Crossover probability , is represented as:

[0061] In the formula This is the set lower limit value for the difference coefficient of variation; This is the upper limit value set for the coefficient of variation. This is the upper limit of the set crossover probability; This is the set lower limit value for the crossover probability;

[0062] Construct neighborhood guiding vectors:

[0063] Construct the neighborhood guiding vector of agent a , is represented as:

[0064] In the formula This is the k-th generation local best individual of agent a; Let a be the set of neighboring smart agents;

[0065] Difference variation and crossover:

[0066] The decision vector for the p-th individual in the k-th generation Construct the difference mutation vector, then the mutated individuals Represented as:

[0067] In the formula For in set A randomly selected index that is different from p; The neighborhood guidance weight coefficients for the defined agent a;

[0068] Binomial crossover was used to generate experimental individuals. ;for The j-th dimension component , is represented as:

[0069] In the formula For range A random number that is uniformly distributed on the upper surface; for Random dimension index on; Let be the j-th dimension component of agent a in the p-th mutated individual of the k-th generation; Let be the j-th dimension component of agent a in the p-th target individual of the k-th generation;

[0070] Fitness assessment and population renewal:

[0071] For test individuals By combining with other agents' current reference individuals, a global candidate solution is formed. Original target individual The global solution formed by the combination is denoted as ; using the augmented local objective function defined in step S4 Perform fitness evaluation:

[0072] If satisfied Then accept the test subjects and let And update the local best individual ;

[0073] Otherwise, retain the original target individual, and let ;in, Let agent a be the target individual in the kth generation and pth position.

[0074] Termination criteria and result output:

[0075] when When the time is right, the algorithm terminates; select the combination of the local best individuals of each agent to form a global decision vector, which is used as the output of the multi-agent differential evolution cooperative self-optimization algorithm;

[0076] Otherwise, the value of k is incremented by 1, and the process returns to the "update adaptive parameters" step to continue iterative evolution until the termination condition is met.

[0077] Step S6 describes solving the multi-agent structure optimization scheduling model using the method designed in step S5, specifically including the following steps:

[0078] The control variables are uniformly represented as a vector of decision variables. ;

[0079] According to the multi-agent division in step S4, Decomposed into a set of local decision variables for each agent. ;

[0080] Based on the objective function constructed in step S3, and combined with the constraints in step S2, the optimization model is obtained, expressed as:

[0081] Following step S4, the optimization model is reconstructed into a multi-agent joint optimization form, expressed as:

[0082] The model is solved using step S4 to obtain the final solution result;

[0083] The solution results include the power purchase and sale curves of the distribution network and the upper-level power grid during the scheduling cycle, the power curves of each microgrid interconnection line, the wind power and photovoltaic power output allocation scheme, the energy storage charging and discharging plan, and the flexible load adjustment trajectory.

[0084] This invention also provides a system for implementing the aforementioned energy optimization scheduling method based on distribution network-microgrid collaboration, comprising a data acquisition module, a model building module, a target building module, a scheduling building module, a solution building module, and an optimization scheduling module; the data acquisition module, model building module, target building module, scheduling building module, solution building module, and optimization scheduling module are connected in series; the data acquisition module is used to acquire data information of the target distribution network and microgrid, and upload the data information to the model building module; the model building module is used to construct a mathematical model and constraints of the distribution network-microgrid collaborative operation system based on the received data information and the acquired data information, and upload the data information to the target building module; the target building module is used to construct a mathematical model and constraints of the distribution network-microgrid collaborative operation system based on the received data information. Based on the constructed model and constraints, an energy optimization scheduling objective function based on distribution network-microgrid collaboration is constructed, and the data information is uploaded to the scheduling construction module. The scheduling construction module is used to construct a multi-agent structure of distribution network-microgrid based on the received data information and the constructed objective function to form a multi-agent structure optimization scheduling model, and the data information is uploaded to the solution construction module. The solution construction module is used to design a self-optimization algorithm based on multi-agent differential evolution for the obtained model based on the received data information, and the data information is uploaded to the optimization scheduling module. The optimization scheduling module is used to solve the multi-agent structure optimization scheduling model using the designed method based on the received data information, and complete the energy optimization scheduling based on distribution network-microgrid collaboration.

[0085] The energy optimization scheduling method and system based on distribution network-microgrid collaboration provided by this invention acquires and processes information from the target distribution network and microgrid, constructs a corresponding optimization scheduling model, and performs model processing, transformation, and solution. This not only achieves energy optimization scheduling based on distribution network-microgrid collaboration, but also has higher reliability, better accuracy, and higher efficiency. Attached Figure Description

[0086] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0087] Figure 2 This is a schematic diagram of the functional modules of the system of the present invention. Detailed Implementation

[0088] like Figure 1 The diagram shown illustrates the method flow of this invention: This energy optimization scheduling method based on distribution network-microgrid collaboration, disclosed in this invention, includes the following steps:

[0089] S1. Obtain data information from the target distribution network and microgrid; specifically including the following steps:

[0090] Acquire data information from the target distribution network and microgrid;

[0091] The data information includes parameter data information of distributed power sources, operation data information of distributed power sources, operation data information of distribution networks and microgrids, and parameter data information of distribution networks and microgrids.

[0092] S2. Based on the data obtained in step S1, construct the mathematical model and constraints of the distribution network-microgrid collaborative operation system; specifically including the following steps:

[0093] This step establishes a mathematical model of the collaborative operation system consisting of the distribution network and multiple microgrids under the grid-connected operation mode, and performs unified modeling of network topology, distributed power sources, energy storage, small generating units, flexible loads, and tie lines;

[0094] The node power balance model is constructed using the following formula:

[0095] In the formula Let be the equivalent active load of node i at time t; For the set of distributed power sources connected to node i; Let g be the active power output of the distributed power source g at time t; Let be the energy storage discharge power of node i at time t; Let be the energy storage charging power of node i at time t; The active power exchange between node i and the upstream power grid or tie line at time t; Let i be the equivalent load at time t after considering flexible adjustment. Let be the equivalent reactive load of node i at time t; Let g be the reactive power output of the distributed power source g at time t. Let be the energy storage reactive power of node i at time t; The reactive power interaction between node i and the upstream power grid or tie line at time t; Let be the voltage amplitude at node i at time t; A set of nodes within the system; Let be the voltage amplitude at node j at time t; branch road The real part of the admittance; branch road The imaginary part of the admittance; Let be the voltage phase angle difference between node i and node j at time t;

[0096] The following formula is used as the active power output constraint for distributed generation:

[0097] In the formula Let g be the lower limit of the active power output of the distributed power source g at time t. Let g be the grid-connected state variable of the distributed power source g. If the distributed power source g is in the grid-connected state at time t, then... If the distributed power source g is in a non-grid-connected state at time t, then ; Let g be the active power output of the distributed power source g at time t; Let g be the upper limit of the active power output of the distributed power source g at time t; Let g be the reactive power output of the distributed power source g at time t. The rated capacity of the distributed power source g;

[0098] If the distributed power source g is a gas turbine, then the following additional ramp constraint is added:

[0099] In the formula The ramp-up limit for distributed power source g is reduced; The ramp-up limit for distributed power source g;

[0100] If the distributed power source g is distributed photovoltaic (PV), then the following formula is used as the output capacity constraint of the PV inverter:

[0101] In the formula The photovoltaic active power available for distributed power source g at time t can be estimated from irradiance and temperature. The rated apparent capacity of the photovoltaic inverter corresponding to the distributed power source g; This is the voltage-reactive power droop factor; This is a reference value for the node voltage; This represents the lower limit of reactive power output of the photovoltaic inverter. This represents the upper limit of reactive power output of the photovoltaic inverter. It is a saturation function, and This function is used to truncate the linear droop relationship at... Within the range;

[0102] For energy storage devices, the following formula is used as the energy storage system model:

[0103] In the formula Let be the maximum charging power of the energy storage device at node i at time t; Let be the charging state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a charging state at time t, then... If the energy storage device at node i is in a non-charging state at time t, then ; Let be the maximum discharge power of the energy storage device at node i at time t; Let be the discharge state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a discharge state at time t, then... If the energy storage device at node i is in a non-discharge state at time t, then ; Let t be the energy state of the energy storage device at time t; This refers to the charging power. This refers to the discharge power. To adjust the step size; This is the lower limit of the energy of the energy storage device; This represents the upper limit of the energy of the energy storage device. ;

[0104] For energy storage devices, the following formula is used as the reactive power constraint for energy storage:

[0105] In the formula Let be the net active power of the energy storage device at node i at time t, and ; Let be the rated apparent capacity of the energy storage device converter at node i;

[0106] For the flexible load at node i, the following formula is used as the flexible load model and constraint:

[0107] In the formula Let be the baseline load of node i at time t; Let be the adjustment amount at node i at time t; Let i be the load reduction limit for node i; Set the load increase limit for node i;

[0108] For transferable loads, the formula is used. As an energy conservation constraint within a given scheduling period T;

[0109] The following formula is used as the tie line and operational safety constraint:

[0110] In the formula Let be the active power of the tie line between microgrid m and the upper-level distribution network at time t; Let be the reactive power of the tie line between microgrid m and the upper-level distribution network at time t; The capacity of the tie line between the microgrid m and the upper-level distribution network; Let be the lower limit of the voltage at node i; This represents the upper limit of the voltage at node i; branch at time t The active power; branch at time t reactive power; branch road Maximum capacity;

[0111] S3. Based on the model and constraints established in step S2, construct an energy optimization scheduling objective function based on distribution network-microgrid coordination; specifically including the following steps:

[0112] Based on the physical model and symbol conventions of step S2, this step constructs a safety-economic integrated objective function for collaborative energy management;

[0113] Within the scheduling period T, the following formula is used as the system's full-cycle decision vector. :

[0114] In the formula Let be the system decision vector at time t; Let t be the decision vector of the distribution network coordination layer; Let be the local decision vector of microgrid m at time t; For the system's microgrid collection; For a microgrid m, it is a collection of distributed power sources. A collection of energy storage devices within a microgrid m; The set of flexible load nodes within a microgrid m; This represents the power purchased by the distribution network and the upstream power grid at time t. This represents the electricity sales power of the distribution network and the upstream power grid at time t. The energy state of energy storage device s within microgrid m at time t;

[0115] With the goal of minimizing the total operating cost within the scheduling cycle, and introducing penalty terms for node voltage exceeding limits and branch power flow exceeding limits, the following formula is adopted as the objective function for energy optimization scheduling based on distribution network-microgrid collaboration. :

[0116] In the formula The cost of purchasing and selling electricity from the upper-level power grid, and , The electricity purchase price between the distribution network and the upstream power grid at time t. The electricity price at time t represents the price at which the distribution network and the upstream power grid are sold. For the operating cost of distributed power sources, and , The first cost coefficient for secondary fuels The second cost coefficient for secondary fuels. This is the third cost factor for secondary fuels; for purely renewable wind and solar power, it can make... ; The startup cost for distributed power sources. Let be the startup state variable of the distributed power source at time t. If the distributed power source is in the startup state at time t, then... If the distributed power source is in a non-startup state at time t, then... , The cost of downtime for distributed power sources. Let be the shutdown state variable of the distributed power source at time t. If the distributed power source is in a shutdown state at time t, then... If the distributed power source is in a non-shutdown state at time t, then , ; The costs associated with energy storage charging, discharging, and depreciation, and , Let be the charging power of the energy stored at time t, s. Let be the discharge power of the stored energy s at time t. The unit charging power depreciation cost factor for energy storage s (representing the lifespan loss caused by cycling). The unit discharge cost factor for energy storage s (representing the lifespan loss caused by cycling); For flexible load adjustment costs, and , This represents the upward displacement relative to the baseline load (indicating an increase in electricity consumption). This represents a reduction in load relative to the baseline load (indicating a cut in electricity consumption). Let i be the unit positive adjustment cost. The unit negative adjustment cost for node i; For network loss costs, and , For grid loss electricity price, For system network loss, , For the collection of distribution network branches, For microgrid branch collection, branch road The active power loss at time t, , branch road The equivalent resistance, branch road The meritorious trend, branch road The unproductive current; Weighting for voltage over-limit penalties; Let be the voltage over-limit penalty function, and , For positive part operators, , It is the square of the positive part operator (used to increase the penalty for exceeding the limit). Weighting of branch circuit overload penalties; Let be the branch overload penalty function, and ; and Used to balance economy and safety;

[0117] S4. Based on the objective function constructed in step S3, construct a multi-agent structure for the distribution network-microgrid to form an optimal scheduling model for the multi-agent structure; specifically, this includes the following steps:

[0118] The agents are divided into groups, and the set of agents is represented as follows:

[0119] In the formula For distribution network coordination intelligent agents; For microgrid intelligent agents;

[0120] The following formula is used as Local cost function :

[0121] The following formula is used as Local cost function :

[0122] At this point, the global objective function Represented as ;

[0123] The tie-line power residual is introduced to characterize the boundary power consistency between the distribution network and each microgrid; for microgrid m, the active power injected at time t by the tie node is calculated according to the power flow equation. and injected reactive power Then the tie line power residual Represented as:

[0124] In the formula The active power residual of the tie line; For the reactive power residual of the tie line; if and only if At that time, the tie line power is the same as the injection power;

[0125] Constructing the tie-line power consistency penalty function for microgrid m , represented as ,in The square of the L2 norm;

[0126] The tie-line power consistency penalty term is distributed to the local objective function of each agent, and the augmented local objective function of each agent is set as follows:

[0127] In the formula For intelligent agents The augmented local objective function; This is the first non-negative weighting coefficient; For intelligent agents The augmented local objective function; This is the second non-negative weighting coefficient; and Used to adjust the degree of importance that each agent attaches to the consistency of the tie-line power;

[0128] Finally, the multi-agent joint optimization model for micro-cooperative operation is expressed as:

[0129] In the formula The feasible region is formed by the various constraints in step S2;

[0130] According to the principle of the quadratic penalty function method, when the penalty factor is sufficiently large, the solution of the multi-agent joint optimization model can approximate the original constrained optimization problem within the feasible region.

[0131] S5. For the model obtained in step S4, design a self-optimization algorithm based on multi-agent differential evolution; specifically including the following steps:

[0132] This step presents the algorithm for solving step S4, the multi-agent differential evolution algorithm. The algorithm uses... To improve fitness, parameters are adaptively adjusted based on operational safety indicators, and neighborhood optimal solutions are used to guide the population of agents to converge collaboratively toward a global solution with consistent power and operational safety.

[0133] Algorithm initialization:

[0134] Set the evolution iteration number index k takes the value For any intelligent agent Set the population size as The local decision vector of the p-th individual in the k-th generation is represented as: ;

[0135] For any algebra The global reference solution is formed by combining the best local individuals of each agent. It is used to measure the operational safety level of this generation and as a benchmark for constructing global candidate solutions;

[0136] Update adaptive parameters:

[0137] Constructing a normalized index based on global operational safety , is represented as:

[0138] In the formula This is the total global security penalty for the kth generation calculated according to step S3; This represents the total global security penalty corresponding to the 0th generation population. This is a set minimum value to prevent the denominator from being 0;

[0139] Define adaptive difference coefficient of variation Crossover probability , is represented as:

[0140] In the formula This is the set lower limit value for the difference coefficient of variation; This is the upper limit value set for the coefficient of variation. This is the upper limit of the set crossover probability; This is the set lower limit for the crossover probability; when the system is operating with significant voltage exceedance or overload, Larger near , near The variable asynchronous length is increased, and the crossover is more aggressive to enhance global exploration; when the system operation basically satisfies the constraints, Smaller near , near By reducing the variable asynchronous length and adopting a conservative cross-biased approach, the self-optimizing characteristic of adaptive convergence towards a safe solution is achieved.

[0141] Construct neighborhood guiding vectors:

[0142] Construct the neighborhood guiding vector of agent a , is represented as:

[0143] In the formula This is the k-th generation local best individual of agent a; Let a be the set of neighboring smart agents;

[0144] Difference variation and crossover:

[0145] The decision vector for the p-th individual in the k-th generation Construct the difference mutation vector, then the mutated individuals Represented as:

[0146] In the formula For in set A randomly selected index that is different from p; The neighborhood guidance weight coefficients for the defined agent a;

[0147] Binomial crossover was used to generate experimental individuals. ;for The j-th dimension component , is represented as:

[0148] In the formula For range A random number that is uniformly distributed on the upper surface; for Random dimension index on; Let be the j-th dimension component of agent a in the p-th mutated individual of the k-th generation; Let be the j-th dimension component of agent a in the p-th target individual of the k-th generation;

[0149] Fitness assessment and population renewal:

[0150] For test individuals By combining with other agents' current reference individuals, a global candidate solution is formed. Original target individual The global solution formed by the combination is denoted as ; using the augmented local objective function defined in step S4 Perform fitness evaluation:

[0151] If satisfied Then accept the test subjects and let And update the local best individual ;

[0152] Otherwise, retain the original target individual, and let ;in, Let agent a be the target individual in the kth generation and pth position.

[0153] Termination criteria and result output:

[0154] when When the time is right, the algorithm terminates; select the combination of the local best individuals of each agent to form a global decision vector, which is used as the output of the multi-agent differential evolution cooperative self-optimization algorithm;

[0155] Otherwise, the value of k is incremented by 1, and the process returns to the "update adaptive parameters" step to continue iterative evolution until the termination condition is met;

[0156] S6. Using the method designed in step S5, solve the multi-agent structure optimization scheduling model to complete the energy optimization scheduling based on distribution network-microgrid collaboration; specifically including the following steps:

[0157] The control variables are uniformly represented as a vector of decision variables. ;

[0158] According to the multi-agent division in step S4, Decomposed into a set of local decision variables for each agent. ;

[0159] Based on the objective function constructed in step S3, and combined with the constraints in step S2, the optimization model is obtained, expressed as:

[0160] Following step S4, the optimization model is reconstructed into a multi-agent joint optimization form, expressed as:

[0161] The model is solved using step S4 to obtain the final solution result;

[0162] The solution results include the power purchase and sale curves of the distribution network and the upper-level grid during the scheduling period, the power curves of each microgrid interconnection line, the wind power and photovoltaic power output allocation scheme, the energy storage charging and discharging plan, and the flexible load adjustment trajectory, which can be directly used as the distribution-microgrid coordinated energy optimization scheduling scheme under the grid-connected operation mode.

[0163] like Figure 2 The diagram shows the functional modules of the system of the present invention: The system disclosed in this invention for implementing the energy optimization scheduling method based on distribution network-microgrid collaboration includes a data acquisition module, a model building module, a target building module, a scheduling building module, a solution building module, and an optimization scheduling module; these modules are connected in series. The data acquisition module acquires data information from the target distribution network and microgrid, and uploads the data information to the model building module. The model building module constructs a mathematical model and constraints for the distribution network-microgrid collaborative operation system based on the received and acquired data information, and uploads the data information to the target building module. The target building module is used to... Based on the received data, and according to the constructed model and constraints, an energy optimization scheduling objective function based on distribution network-microgrid collaboration is constructed, and the data is uploaded to the scheduling construction module. The scheduling construction module, based on the received data and the constructed objective function, constructs a multi-agent structure for the distribution network-microgrid to form a multi-agent structure optimization scheduling model, and uploads the data to the solution construction module. The solution construction module, based on the received data and the obtained model, designs a self-optimizing algorithm based on multi-agent differential evolution, and uploads the data to the optimization scheduling module. The optimization scheduling module, based on the received data and using the designed method, solves the multi-agent structure optimization scheduling model to complete the energy optimization scheduling based on distribution network-microgrid collaboration.

Claims

1. An energy optimization scheduling method based on distribution network-microgrid coordination, comprising the following steps: S1. Obtain data information from the target distribution network and microgrid; S2. Based on the data obtained in step S1, construct the mathematical model and constraints of the distribution network-microgrid collaborative operation system; S3. Based on the model and constraints established in step S2, construct an energy optimization scheduling objective function based on the coordination of distribution network and microgrid; S4. Based on the objective function constructed in step S3, construct a multi-agent structure for the distribution network-microgrid to form an optimized scheduling model for the multi-agent structure; S5. For the model obtained in step S4, design a self-optimization algorithm based on multi-agent differential evolution; S6. Using the method designed in step S5, solve the multi-agent structure optimization scheduling model to complete the energy optimization scheduling based on the coordination of distribution network and microgrid.

2. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 1, characterized in that... Step S1, which involves acquiring data information about the target distribution network and microgrid, specifically includes the following steps: Acquire data information from the target distribution network and microgrid; The data information includes parameter data information of distributed power sources, operation data information of distributed power sources, operation data information of distribution networks and microgrids, and parameter data information of distribution networks and microgrids.

3. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 2, characterized in that... Step S2, which involves constructing a mathematical model and constraints for the distribution network-microgrid collaborative operation system based on the data information obtained in step S1, specifically includes the following steps: The node power balance model is constructed using the following formula: In the formula Let be the equivalent active load of node i at time t; For the set of distributed power sources connected to node i; Let g be the active power output of the distributed power source g at time t; Let be the energy storage discharge power of node i at time t; Let be the energy storage charging power of node i at time t; The active power exchange between node i and the upstream power grid or tie line at time t; Let i be the equivalent load at time t after considering flexible adjustment. Let be the equivalent reactive load of node i at time t; Let g be the reactive power output of the distributed power source g at time t. Let be the energy storage reactive power of node i at time t; The reactive power interaction between node i and the upstream power grid or tie line at time t; Let be the voltage amplitude at node i at time t; A set of nodes within the system; Let be the voltage amplitude at node j at time t; branch road The real part of the admittance; branch road The imaginary part of the admittance; Let be the voltage phase angle difference between node i and node j at time t; The following formula is used as the active power output constraint for distributed generation: In the formula Let g be the lower limit of the active power output of the distributed power source g at time t. Let g be the grid-connected state variable of the distributed power source g. If the distributed power source g is in the grid-connected state at time t, then... If the distributed power source g is in a non-grid-connected state at time t, then ; Let g be the active power output of the distributed power source g at time t; Let g be the upper limit of the active power output of the distributed power source g at time t; Let g be the reactive power output of the distributed power source g at time t. The rated capacity of the distributed power source g; If the distributed power source g is a gas turbine, then the following additional ramp constraint is added: In the formula The ramp-up limit for distributed power source g is reduced; The ramp-up limit for distributed power source g; If the distributed power source g is distributed photovoltaic (PV), then the following formula is used as the output capacity constraint of the PV inverter: In the formula Let be the photovoltaic active power available from distributed power source g at time t; The rated apparent capacity of the photovoltaic inverter corresponding to the distributed power source g; This is the voltage-reactive power droop factor; This is a reference value for the node voltage; This represents the lower limit of reactive power output of the photovoltaic inverter. This represents the upper limit of reactive power output of the photovoltaic inverter. It is a saturation function, and ; For energy storage devices, the following formula is used as the energy storage system model: In the formula Let be the maximum charging power of the energy storage device at node i at time t; Let be the charging state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a charging state at time t, then... If the energy storage device at node i is in a non-charging state at time t, then ; Let be the maximum discharge power of the energy storage device at node i at time t; Let be the discharge state variable of the energy storage device at node i at time t. If the energy storage device at node i is in a discharge state at time t, then... If the energy storage device at node i is in a non-discharge state at time t, then ; Let t be the energy state of the energy storage device at time t; This refers to the charging power. This refers to the discharge power. To adjust the step size; This is the lower limit of the energy of the energy storage device; This represents the upper limit of the energy of the energy storage device. For energy storage devices, the following formula is used as the reactive power constraint for energy storage: In the formula Let be the net active power of the energy storage device at node i at time t, and ; Let be the rated apparent capacity of the energy storage device converter at node i; For the flexible load at node i, the following formula is used as the flexible load model and constraint: In the formula Let be the baseline load of node i at time t; Let be the adjustment amount at node i at time t; Let i be the load reduction limit for node i; Set the load increase limit for node i; For transferable loads, the formula is used. As an energy conservation constraint within a given scheduling period T; The following formula is used as the tie line and operational safety constraint: In the formula Let be the active power of the tie line between microgrid m and the upper-level distribution network at time t; Let be the reactive power of the tie line between microgrid m and the upper-level distribution network at time t; The capacity of the tie line between the microgrid m and the upper-level distribution network; Let be the lower limit of the voltage at node i; This represents the upper limit of the voltage at node i; branch at time t The active power; branch at time t reactive power; branch road Maximum capacity.

4. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 3, characterized in that... Step S3, which involves constructing an energy optimization scheduling objective function based on the model and constraints established in step S2, specifically includes the following steps: Within the scheduling period T, the following formula is used as the system's full-cycle decision vector. : In the formula Let be the system decision vector at time t; Let t be the decision vector of the distribution network coordination layer; Let be the local decision vector of microgrid m at time t; For the system's microgrid collection; For a microgrid m, it is a collection of distributed power sources. A collection of energy storage devices within a microgrid m; The set of flexible load nodes within a microgrid m; This represents the power purchased by the distribution network and the upstream power grid at time t. This represents the electricity sales power of the distribution network and the upstream power grid at time t. The energy state of energy storage device s within microgrid m at time t; The following formula is adopted as the objective function for energy optimization scheduling based on distribution network-microgrid coordination. : In the formula The cost of purchasing and selling electricity from the upper-level power grid, and , The electricity purchase price between the distribution network and the upstream power grid at time t. The electricity price at time t represents the price at which the distribution network and the upstream power grid are sold. For the operating cost of distributed power sources, and , The first cost coefficient for secondary fuels The second cost coefficient for secondary fuels. The third cost coefficient for secondary fuels; The startup cost for distributed power sources. Let be the startup state variable of the distributed power source at time t. If the distributed power source is in the startup state at time t, then... If the distributed power source is in a non-startup state at time t, then... , The cost of downtime for distributed power sources. Let be the shutdown state variable of the distributed power source at time t. If the distributed power source is in a shutdown state at time t, then... If the distributed power source is in a non-shutdown state at time t, then , ; The costs associated with energy storage charging, discharging, and depreciation, and , Let be the charging power of the energy stored at time t, s. Let be the discharge power of the stored energy s at time t. The depreciation cost factor per unit charging capacity of energy storage s. The unit discharge capacity loss cost factor for energy storage s; For flexible load adjustment costs, and , This is the upward displacement relative to the reference load. This is the amount of reduction relative to the baseline load. Let i be the unit positive adjustment cost. The unit negative adjustment cost for node i; For network loss costs, and , For grid loss electricity price, For system network loss, , For the collection of distribution network branches, For microgrid branch collection, branch road The active power loss at time t, , branch road The equivalent resistance, branch road The meritorious trend, branch road The unproductive current; Weighting for voltage over-limit penalties; Let be the voltage over-limit penalty function, and , For positive part operators, , The square of the positive part operator; Weighting of branch circuit overload penalties; Let be the branch overload penalty function, and .

5. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 4, characterized in that... Step S4, which involves constructing a distribution network-microgrid multi-agent structure based on the objective function built in step S3 to form a multi-agent structure optimization scheduling model, specifically includes the following steps: The agents are divided into groups, and the set of agents is represented as follows: In the formula For distribution network coordination intelligent agents; For microgrid intelligent agents; The following formula is used as Local cost function : The following formula is used as Local cost function : At this point, the global objective function Represented as ; For a microgrid m, the active power injected at time t by the tie nodes, calculated according to the power flow equations, is... and injected reactive power Then the tie line power residual Represented as: In the formula The active power residual of the tie line; For the reactive power residual of the tie line; if and only if At that time, the tie line power is the same as the injection power; Constructing the tie-line power consistency penalty function for microgrid m , represented as ,in The square of the L2 norm; The tie-line power consistency penalty term is distributed to the local objective function of each agent, and the augmented local objective function of each agent is set as follows: In the formula For intelligent agents The augmented local objective function; This is the first non-negative weighting coefficient; For intelligent agents The augmented local objective function; This is the second non-negative weighting coefficient; Finally, the multi-agent joint optimization model for micro-cooperative operation is expressed as: In the formula This is the feasible region formed by the various constraints in step S2.

6. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 5, characterized in that... Step S5 describes designing a self-optimization algorithm based on multi-agent differential evolution for the model obtained in step S4, specifically including the following steps: Algorithm initialization: Set the evolution iteration number index k takes the value For any intelligent agent Set the population size as The local decision vector of the p-th individual in the k-th generation is represented as: ; For any algebra The global reference solution is formed by combining the best local individuals of each agent. ; Update adaptive parameters: Constructing a normalized index based on global operational safety , is represented as: In the formula This is the total global security penalty for the kth generation calculated according to step S3; This represents the total global security penalty corresponding to the 0th generation population. This is a set minimum value to prevent the denominator from being 0; Define adaptive difference coefficient of variation Crossover probability , is represented as: In the formula This is the set lower limit value for the difference coefficient of variation; This is the upper limit value set for the coefficient of variation. This is the upper limit of the set crossover probability; This is the set lower limit value for the crossover probability; Construct neighborhood guiding vectors: Construct the neighborhood guiding vector of agent a , is represented as: In the formula This is the k-th generation local best individual of agent a; Let a be the set of neighboring smart agents; Difference variation and crossover: The decision vector for the p-th individual in the k-th generation Construct the difference mutation vector, then the mutated individuals Represented as: In the formula For in set A randomly selected index that is different from p; The neighborhood guidance weight coefficients for the defined agent a; Binomial crossover was used to generate experimental individuals. ;for The j-th dimension component , is represented as: In the formula For range A random number that is uniformly distributed on the upper surface; for Random dimension index on; Let be the j-th dimension component of agent a in the p-th mutated individual of the k-th generation; Let be the j-th dimension component of agent a in the p-th target individual of the k-th generation; Fitness assessment and population renewal: For test individuals By combining with other agents' current reference individuals, a global candidate solution is formed. Original target individual The global solution formed by the combination is denoted as ; Using the augmented local objective function defined in step S4 Perform fitness evaluation: If satisfied Then accept the test subjects and let And update the local best individual ; Otherwise, retain the original target individual, and let ;in, Let agent a be the target individual in the kth generation and pth position. Termination criteria and result output: when When the time is right, the algorithm terminates; select the combination of the local best individuals of each agent to form a global decision vector, which is used as the output of the multi-agent differential evolution cooperative self-optimization algorithm; Otherwise, the value of k is incremented by 1, and the process returns to the "update adaptive parameters" step to continue iterative evolution until the termination condition is met.

7. The energy optimization scheduling method based on distribution network-microgrid coordination according to claim 6, characterized in that... Step S6 describes solving the multi-agent structure optimization scheduling model using the method designed in step S5, specifically including the following steps: The control variables are uniformly represented as a vector of decision variables. ; According to the multi-agent division in step S4, Decomposed into a set of local decision variables for each agent. ; Based on the objective function constructed in step S3, and combined with the constraints in step S2, the optimization model is obtained, expressed as: Following step S4, the optimization model is reconstructed into a multi-agent joint optimization form, expressed as: The model is solved using step S4 to obtain the final solution result; The solution results include the power purchase and sale curves of the distribution network and the upper-level power grid during the scheduling cycle, the power curves of each microgrid interconnection line, the wind power and photovoltaic power output allocation scheme, the energy storage charging and discharging plan, and the flexible load adjustment trajectory.

8. A system for implementing the energy optimization scheduling method based on distribution network-microgrid coordination as described in any one of claims 1 to 7, characterized in that... It includes a data acquisition module, a model building module, a target building module, a scheduling building module, a solution building module, and an optimization scheduling module; these modules are connected in series. The data acquisition module acquires data information from the target distribution network and microgrid and uploads the data information to the model building module. The model building module constructs a mathematical model and constraints for the distribution network-microgrid collaborative operation system based on the received and acquired data information, and uploads the data information to the target building module. The objective construction module is used to construct an energy optimization scheduling objective function based on the received data, the constructed model, and constraints, and upload the data to the scheduling construction module. The scheduling construction module is used to construct a multi-agent structure of the distribution network and microgrid based on the received data and the constructed objective function, to form a multi-agent structure optimization scheduling model, and upload the data to the solution construction module. The solution construction module is used to design a self-optimizing algorithm based on multi-agent differential evolution for the obtained model based on the received data, and upload the data to the optimization scheduling module. The optimization scheduling module is used to solve the multi-agent structure optimization scheduling model based on the received data information and the designed method, so as to complete the energy optimization scheduling based on the coordination of distribution network and microgrid.

Images