Flexible optimization method for flexible supply and demand resource collaboration of active power distribution network

Abstract

The invention discloses an elastic optimization method for flexible supply and demand resource cooperation of an active power distribution network. The method comprises the following steps: modeling the uncertainty of distributed power supply output based on meteorological data; representing the uncertainty of load prediction through a load probability density function based on normal distribution; establishing an active power distribution network flexible supply and demand resource coordinated operation optimization model considering the uncertainty of the flexible supply and demand resources; aiming at the problem of uncertainty of flexible supply and demand resources, converting the flexible supply and demand resources into a deterministic expected value model; solving by using a double-layer combinatorial optimization algorithm; according to the method, controllability and uncertainty of supply and demand resources are considered, and the active power distribution network is optimized by taking reduction of network loss as a target.

Description

Technical Field

[0001] The present invention belongs to the technical field of elastic optimization of active distribution networks, and specifically relates to an elastic optimization method for coordinated supply and demand resources of an active distribution network. Background Art

[0002] Distributed renewable energy, energy storage and microgrids have changed the characteristics of user electricity consumption, making the traditional radial feeder distribution network gradually develop into an active distribution network with flexible supply and demand resources; among them, the application of technologies such as distributed power sources, electric energy storage and integrated energy has promoted the diversification of energy sources in the distribution network. At the distribution network level, the application of power electronics in the primary electrical network, the full coverage of the secondary information network and the large-scale emergence of new load terminals such as smart homes, electric vehicles and integrated energy on the load side have made the management and control of the power system more difficult; at the same time, as the penetration rate of distributed energy in the distribution network continues to increase, its intermittent, volatility and uncertainty have changed the operation mode of power systems such as transmission networks and distribution networks, while also increasing the safety risks of distribution networks.

[0003] Therefore, in order to solve the above problems, it is necessary to develop a flexible optimization method for the coordination of flexible supply and demand resources in active distribution networks. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and to provide a flexible optimization method for the coordination of supply and demand resources of an active distribution network, taking into account the controllability and uncertainty of supply and demand resources, and optimizing the active distribution network with the goal of reducing network losses.

[0005] The object of the present invention is achieved by: a method for elastic optimization of flexible supply and demand resource coordination of active distribution network, characterized in that it comprises the following steps:

[0006] S1. Model the uncertainty of distributed power output based on meteorological data; including the uncertainty brought by the access of wind power and photovoltaic power to the distribution network, describing the uncertainty of wind power output through the probability density distribution function based on Weibull distribution, and describing the uncertainty of photovoltaic power output through the probability density distribution function based on Beta distribution;

[0007] S2, the uncertainty of load forecasting is expressed by the load probability density function based on normal distribution;

[0008] S3. Establish an optimization model for the coordinated operation of flexible supply and demand resources in active distribution networks taking into account the uncertainty of flexible supply and demand resources;

[0009] S31. Taking the minimum network loss value of the active distribution network as the objective function, its expression is as follows:

[0010]

[0011] In the formula, is the active power injected into the node i at time t, which is numerically equal to the network loss power at that time, and n is the total number of nodes;

[0012] S32, constructing constraint conditions of the objective function; the constraint conditions include topology constraint, state of charge constraint, power balance constraint, node voltage constraint, branch current constraint and predicted load constraint;

[0013] S4. In view of the uncertainty of flexible supply and demand resources, transform them into a deterministic expected value model; specifically, it includes the uncertainty transformation of distributed power sources and the uncertainty transformation of loads;

[0014] S5. Use a two-layer combined optimization algorithm to solve the problem. By changing the status of the segments and interconnecting switches one by one and satisfying various constraints, a nonlinear programming method is used to calculate the local optimal target values ​​under different distribution network topologies to achieve lower-level optimization. Then, different local optimal solutions are compared to obtain the best local optimal solution to achieve upper-level optimization.

[0015] Furthermore, the description of the uncertainty of the wind turbine output by the probability density distribution function based on the Weibull distribution in step S1 specifically includes:

[0016] The probability distribution of wind speed follows the Weibull distribution, and its probability density function is:

[0017]

[0018] In the formula, v is the actual wind speed, c is the scale parameter, and k is the shape parameter;

[0019] Then its cumulative distribution function is:

[0020]

[0021] Furthermore, the description of the uncertainty of the output of the photovoltaic unit by the probability density distribution function based on the Beta distribution in step S1 specifically includes:

[0022] Since photovoltaic output power is mainly affected by solar irradiance, the solar irradiance in a certain period of time is approximately Beta distributed, and its probability density function is as follows:

[0023]

[0024] Where α, β are shape parameters, s is the solar irradiance, Γ is the gamma function, where α≥0, β≥0, 0≤s≤1;

[0025] Based on the photoelectric conversion characteristics of photovoltaic power generation, the output of photovoltaic units also obeys the Beta distribution, and we get:

[0026]

[0027] , where P M For photovoltaic power generation, P max The maximum value of photovoltaic power generation output, f(P M ) is the photovoltaic power generation output P M The corresponding probability density, F(P M ) is the photovoltaic power generation output between 0 and P M The probability between .

[0028] Furthermore, in step S2, the uncertainty of load forecasting is described by a normal distribution model, which is specifically expressed as follows:

[0029]

[0030] In the formula, represents the active load prediction error at node i at time t, ξ T Represents the normal distribution standard deviation of the forecast error in time period T, η T Represents the expected value of the normal distribution of the forecast error in time period T.

[0031] Furthermore, in step S32, in order to keep the topology of the distribution network radial and without an isolated ring network, the topology constraint is expressed as follows:

[0032]

[0033] Where y is the set of substation nodes in the distribution network.

[0034] Furthermore, the state of charge constraint SOC in step S32 is expressed as follows:

[0035]

[0036] Where, E(t) is the remaining energy stored at time t, E ESS,i The energy storage capacity installed for node i;

[0037] S OC,min ≤S OC,i (t)≤S OC,max

[0038] In the formula, S OC,min and S OC,max is the minimum and maximum value of the state of charge pair, where 10% ≤ S OC,min ≤30%, 80%≤S OC,max≤100%;

[0039] The power balance constraint is expressed as follows: the inflow and outflow power of each node satisfies the power balance equation:

[0040]

[0041] In the formula, are the active power and reactive power injected into node i at time t, U i,t , U j,t are the voltage amplitudes of nodes i and j at time t, K ij is the switch connectivity variable between nodes i and j, K ij =1 means the line between nodes i and j remains connected, K ij =0 means the line between nodes i and j is disconnected, G ij , B ij , δ ij They are the conductance, susceptance and voltage phase difference between nodes i and j respectively, and n is the total number of system nodes.

[0042] Furthermore, the node voltage constraint in step S32 is expressed as follows: the voltage of each node cannot exceed its rated upper limit, nor can it be lower than its lower limit:

[0043] U min ≤U i,t ≤U max ,i∈Ω

[0044] Where U min , U max are the lower and upper limits of node voltage respectively, Ω is the set of system nodes;

[0045] The branch current constraint is expressed as follows: the current between nodes i and j cannot exceed its rated upper limit, nor can it be lower than its lower limit:

[0046]

[0047] , where are the lower and upper limits of the node current respectively;

[0048] The forecast load constraint is expressed as, and the forecast load of each node is expressed as follows:

[0049]

[0050] , where represents the predicted active and reactive loads at node i at time t, represents the basic active and reactive loads at node i at time t, Represents the active and reactive load prediction error at node i at time t.

[0051] Furthermore, the uncertainty conversion of the distributed power source in step S4 includes uncertainty conversion of wind turbine output and photovoltaic output.

[0052] Among them, the Weibull distribution expectation of wind turbine output is:

[0053]

[0054] Where Γ is the gamma function; the variance is:

[0055]

[0056] Among them, the Beta distribution expectation of photovoltaic unit output is:

[0057]

[0058] The variance is:

[0059]

[0060] According to the expectation and variance of historical monitoring data combined with the above formula, the parameters α and β are obtained:

[0061]

[0062] Furthermore, in the load uncertainty conversion in step S4, the expected value model of the predicted load is as follows:

[0063]

[0064] In the formula, represents the basic active and reactive loads at node i at time t, represents the active and reactive load prediction error at node i at time t, represents the predicted active and reactive loads at node i at time t, It represents the expected value of active and reactive forecast load at node i at time t.

[0065] Furthermore, the solution using the double-layer combination optimization algorithm in step S5 specifically includes the following steps: ① Input the initial data parameters, set the iteration parameter N A =1; ② Lower-level optimization, using nonlinear programming method to solve the model and obtain the network loss Z; ③ Upper-level optimization, judging whether the network loss obtained in the current iteration is smaller than that in the previous iteration, if so, the first element in the current input switch array is inverted to form a new array K[m], N A=1, repeat step ②; if not, determine whether all elements of array K[m] have been traversed; if so, output the current switch array K[m] and the corresponding local optimal network loss value Z, and end the calculation; if not, let N A =N A +1, traverse the last switch array to the Nth A Each element is inverted to form a new array K[m], and step ② is repeated; ④ double-layer combinatorial optimization is completed; by comparing different local optimal solutions, the best local optimal solution is obtained.

[0066] Due to the adoption of the above technical scheme, the beneficial effects of the present invention are as follows: by modeling the uncertainty of distributed power output and the uncertainty of load forecasting, an optimization model for the coordinated operation of flexible supply and demand resources of the active distribution network is established, and the uncertain problem is processed and converted into a deterministic expected value model, which is solved by a two-layer combined optimization algorithm to optimize the active distribution network, effectively reduce network losses, and achieve dynamic balance of supply and demand. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a flow chart of the present invention.

[0068] Figure 2 It is a flowchart of solving the problem using a double-layer combinatorial optimization algorithm in step S5 of the present invention. DETAILED DESCRIPTION

[0069] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0070] like Figure 1 , Figure 2 As shown, a flexible optimization method for coordinated supply and demand resources of an active distribution network comprises the following steps:

[0071] Step S1, modeling the uncertainty of distributed power output based on meteorological data; including the uncertainty brought by the access of wind power and photovoltaic power to the distribution network, describing the uncertainty of wind power output by a probability density distribution function based on Weibull distribution, and describing the uncertainty of photovoltaic power output by a probability density distribution function based on Beta distribution.

[0072] Specifically, the description of the uncertainty of the wind turbine output by the probability density distribution function based on the Weibull distribution in step S1 specifically includes:

[0073] The probability distribution of wind speed follows the Weibull distribution, and its probability density function is:

[0074]

[0075] In the formula, v is the actual wind speed, c is the scale parameter, and k is the shape parameter;

[0076] Then its cumulative distribution function is:

[0077]

[0078] Specifically, the description of the uncertainty of the photovoltaic unit output by the probability density distribution function based on the Beta distribution in step S1 specifically includes:

[0079] Since the photovoltaic output power is mainly affected by the solar irradiance, the solar irradiance in a certain period of time is approximately Beta distribution, and its probability density function is as follows:

[0080]

[0081] Where α, β are shape parameters, s is the solar irradiance, Γ is the gamma function, where α≥0, β≥0, 0≤s≤1;

[0082] Based on the photoelectric conversion characteristics of photovoltaic power generation, the output of photovoltaic units also obeys the Beta distribution, and we get:

[0083]

[0084] Where P M For photovoltaic power generation, P max The maximum value of photovoltaic power generation output, f(P M ) is the photovoltaic power generation output P M The corresponding probability density, F(P M ) is the photovoltaic power generation output between 0 and P M The probability between .

[0085] Step S2: The uncertainty of load forecasting is represented by a load probability density function based on normal distribution; specifically, the uncertainty of load forecasting is described by a normal distribution model, which is specifically represented as follows:

[0086]

[0087] In the formula, represents the active load prediction error at node i at time t, ξ T Represents the normal distribution standard deviation of the forecast error in time period T, η T Represents the expected value of the normal distribution of the forecast error in time period T.

[0088] Step S3: Establish an active distribution network flexibility supply and demand resource coordinated operation optimization model taking into account the uncertainty of flexibility supply and demand resources.

[0089] Step S31: Taking the minimum network loss value of the active distribution network as the objective function, the expression is as follows:

[0090]

[0091] In the formula, is the active power injected into the node i at time t, which is numerically equal to the network loss power at that time, and n is the total number of nodes.

[0092] Step S32, constructing constraint conditions of the objective function; the constraint conditions include topology constraints, state of charge constraints, power balance constraints, node voltage constraints, branch current constraints and predicted load constraints.

[0093] Specifically, in order to keep the topology of the distribution network radial and without isolated ring networks, the topology constraints are expressed as follows:

[0094]

[0095] Where y is the set of substation nodes in the distribution network.

[0096] Specifically, the state of charge constraint SOC is expressed as follows:

[0097]

[0098] Where, E(t) is the remaining energy stored at time t, E ESS,i The energy storage capacity installed for node i;

[0099] S OC,min ≤S OC,i (t)≤S OC,max

[0100] In the formula, S OC,min and S OC,max is the minimum and maximum value of the state of charge pair, where 10% ≤ S OC,min ≤30%, 80%≤S OC,max ≤100%.

[0101] Specifically, the power balance constraint is expressed as follows: the inflow and outflow power of each node satisfies the power balance equation:

[0102]

[0103] In the formula, are the active power and reactive power injected into node i at time t, U i,t , U j,t are the voltage amplitudes of nodes i and j at time t, K ij is the switch connectivity variable between nodes i and j, Kij =1 means the line between nodes i and j remains connected, K ij =0 means the line between nodes i and j is disconnected, G ij , B ij , δ ij They are the conductance, susceptance and voltage phase difference between nodes i and j respectively, and n is the total number of system nodes.

[0104] Specifically, the node voltage constraint is expressed as follows: the voltage of each node cannot exceed its rated upper limit, nor can it be lower than its lower limit:

[0105] U min ≤U i,t ≤U max ,i∈Ω

[0106] Where U min , U max are the lower and upper limits of node voltage respectively, and Ω is the system node set.

[0107] Specifically, the branch current constraint is expressed as follows: the current between nodes i and j cannot exceed its rated upper limit, nor can it be lower than the lower limit:

[0108]

[0109] In the formula, are the lower and upper limits of the node current respectively.

[0110] Specifically, the forecast load constraint is expressed as, and the forecast load of each node is expressed as follows:

[0111]

[0112] In the formula, represents the predicted active and reactive loads at node i at time t, represents the basic active and reactive loads at node i at time t, Represents the active and reactive load prediction error at node i at time t.

[0113] Step S4: transform the uncertainty of flexible supply and demand resources into a deterministic expected value model, including the uncertainty transformation of distributed power sources and the uncertainty transformation of loads.

[0114] Specifically, the uncertainty transformation of distributed power sources includes the uncertainty transformation of wind turbine output and photovoltaic output, where the Weibull distribution expectation of wind turbine output is:

[0115]

[0116] Where Γ is the gamma function; the variance is:

[0117]

[0118] Among them, the Beta distribution expectation of photovoltaic unit output is:

[0119]

[0120] The variance is:

[0121]

[0122] According to the expectation and variance of historical monitoring data combined with the above formula, the parameters α and β are obtained:

[0123]

[0124] Specifically, in the load uncertainty transformation, the expected value model of the predicted load is as follows:

[0125]

[0126] In the formula, represents the basic active and reactive loads at node i at time t, represents the active and reactive load prediction error at node i at time t, represents the predicted active and reactive loads at node i at time t, It represents the expected value of active and reactive forecast load at node i at time t.

[0127] Step S5, using a two-layer combined optimization algorithm to solve; by changing the state of the segment and the connecting switch one by one and satisfying various constraints, using a nonlinear programming method to calculate the local optimal target value under different distribution network topologies, to achieve lower-level optimization, and then compare different local optimal solutions to obtain the best local optimal solution to achieve upper-level optimization.

[0128] Specifically, the solution using the two-layer combination optimization algorithm includes the following steps: ① Input the initial data parameters, set the iteration parameter N A =1; ② Lower-level optimization, using nonlinear programming method to solve the model and obtain the network loss Z; ③ Upper-level optimization, judging whether the network loss obtained in the current iteration is smaller than that in the previous iteration, if so, the first element in the current input switch array is inverted to form a new array K[m], N A =1, repeat step ②; if not, determine whether all elements of array K[m] have been traversed; if so, output the current switch array K[m] and the corresponding local optimal network loss value Z, and end the calculation; if not, let N A =N A +1, traverse the last switch array to the NthA Each element is inverted to form a new array K[m], and step ② is repeated; ④ double-layer combinatorial optimization is completed; by comparing different local optimal solutions, the best local optimal solution is obtained.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A flexible optimization method for the coordination of supply and demand resources of an active distribution network, characterized by: The following steps are involved: S1. Model the uncertainty of distributed power output based on meteorological data; including the uncertainty brought by the access of wind power and photovoltaic power to the distribution network, describing the uncertainty of wind power output through the probability density distribution function based on Weibull distribution, and describing the uncertainty of photovoltaic power output through the probability density distribution function based on Beta distribution; S2, the uncertainty of load forecasting is expressed by the load probability density function based on normal distribution; S3. Establish an optimization model for the coordinated operation of flexible supply and demand resources in active distribution networks taking into account the uncertainty of flexible supply and demand resources; S31. Taking the minimum network loss value of the active distribution network as the objective function, its expression is as follows: In the formula, is the active power injected into the node i at time t, which is numerically equal to the network loss power at that time, and n is the total number of nodes; S32, constructing constraint conditions of the objective function; the constraint conditions include topology constraint, state of charge constraint, power balance constraint, node voltage constraint, branch current constraint and predicted load constraint; S4. In view of the uncertainty of flexible supply and demand resources, transform them into a deterministic expected value model; specifically, it includes the uncertainty transformation of distributed power sources and the uncertainty transformation of loads; S5. Use a two-layer combined optimization algorithm to solve the problem. By changing the status of the segments and interconnecting switches one by one and satisfying various constraints, a nonlinear programming method is used to calculate the local optimal target values ​​under different distribution network topologies to achieve lower-level optimization. Then, different local optimal solutions are compared to obtain the best local optimal solution to achieve upper-level optimization.

2. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: In step S1, describing the uncertainty of wind turbine output by using a probability density distribution function based on Weibull distribution specifically includes: The probability distribution of wind speed follows the Weibull distribution, and its probability density function is: In the formula, v is the actual wind speed, c is the scale parameter, and k is the shape parameter; Then its cumulative distribution function is:

3. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: In step S1, describing the uncertainty of the output of the photovoltaic unit by using the probability density distribution function based on the Beta distribution specifically includes: Since the photovoltaic output power is mainly affected by the solar irradiance, the solar irradiance in a certain period of time is approximately Beta distribution, and its probability density function is as follows: Where α, β are shape parameters, s is the solar irradiance, Γ is the gamma function, where α≥0, β≥0, 0≤s≤1; Based on the photoelectric conversion characteristics of photovoltaic power generation, the output of photovoltaic units also obeys the Beta distribution, and we get: Where P M For photovoltaic power generation, P max The maximum value of photovoltaic power generation output, f(P M ) is the photovoltaic power generation output P M The corresponding probability density, F(P M ) is the photovoltaic power generation output between 0 and P M The probability between .

4. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1 is characterized by: In step S2, the uncertainty of load forecasting is described by a normal distribution model, which is specifically expressed as: In the formula, represents the active load prediction error at node i at time t, ξ T Represents the normal distribution standard deviation of the forecast error in time period T, η T Represents the expected value of the normal distribution of the forecast error in time period T.

5. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: In step S32, in order to keep the topology of the distribution network radial and without isolated ring networks, the topology constraints are expressed as follows: Where y is the set of substation nodes in the distribution network.

6. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: The state of charge constraint SOC in step S32 is expressed as follows: Where, E(t) is the remaining energy stored at time t, E ESS,i The energy storage capacity installed for node i; S OC,min ≤S OC,i (t)≤S OC,max In the formula, S OC,min and S OC,max is the minimum and maximum value of the state of charge pair, where 10% ≤ S OC,min ≤30%, 80%≤S OC,max ≤100%; The power balance constraint is expressed as follows: the inflow and outflow power of each node satisfies the power balance equation: In the formula, are the active power and reactive power injected into node i at time t, U i,t , U j,t are the voltage amplitudes of nodes i and j at time t, K ij is the switch connectivity variable between nodes i and j, K ij =1 means the line between nodes i and j remains connected, K ij =0 means the line between nodes i and j is disconnected, G ij , B ij , δ ij They are the conductance, susceptance and voltage phase difference between nodes i and j respectively, and n is the total number of system nodes.

7. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: The node voltage constraint in step S32 is expressed as follows: the voltage of each node cannot exceed its rated upper limit, nor can it be lower than its lower limit: IN min ≤U i,t ≤U max ,i∈Ω Where U min , U max are the lower and upper limits of node voltage respectively, Ω is the set of system nodes; The branch current constraint is expressed as follows: the current between nodes i and j cannot exceed its rated upper limit, nor can it be lower than its lower limit: In the formula, are the lower and upper limits of the node current respectively; The forecast load constraint is expressed as, and the forecast load of each node is expressed as follows: In the formula, represents the predicted active and reactive loads at node i at time t, represents the basic active and reactive loads at node i at time t, Represents the active and reactive load prediction error at node i at time t.

8. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: The uncertainty conversion of the distributed power source in step S4 includes the uncertainty conversion of the output of the wind turbine and the output of the photovoltaic unit. Among them, the Weibull distribution expectation of wind turbine output is: Where Γ is the gamma function; the variance is: Among them, the Beta distribution expectation of photovoltaic unit output is: The variance is: According to the expectation and variance of historical monitoring data combined with the above formula, the parameters α and β are obtained:

9. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: In the load uncertainty conversion in step S4, the expected value model of the predicted load is as follows: In the formula, represents the basic active and reactive loads at node i at time t, represents the active and reactive load prediction error at node i at time t, represents the predicted active and reactive loads at node i at time t, It represents the expected value of active and reactive forecast load at node i at time t.

10. The elastic optimization method for the coordination of supply and demand resources of an active distribution network according to claim 1, characterized in that: The step S5 uses a two-layer combination optimization algorithm to solve the problem, which specifically includes the following steps: ① Input the initial data parameters, set the iteration parameter N A =1; ② Lower-level optimization, using nonlinear programming method to solve the model and obtain the network loss Z; ③ Upper-level optimization, judging whether the network loss obtained in the current iteration is smaller than that in the previous iteration, if so, the first element in the current input switch array is inverted to form a new array K[m], N A =1, repeat step ②; if not, determine whether all elements of array K[m] have been traversed; if so, output the current switch array K[m] and the corresponding local optimal network loss value Z, and end the calculation; if not, let N A =N A +1, traverse the last switch array to the Nth A Each element is inverted to form a new array K[m], and step ② is repeated; ④ double-layer combinatorial optimization is completed; by comparing different local optimal solutions, the best local optimal solution is obtained.

Images