Power distribution network SOP operation optimization method and device based on distributed robustness

Abstract

The invention discloses a distribution robustness-based power distribution network SOP operation optimization method and device, and the method comprises the steps: constructing an energy storage SOP model and a relaxation power flow model according to the operation parameter data obtained from a distributed photovoltaic system, and obtaining a first constraint condition and a second constraint condition according to the two models; constructing an initial stochastic optimization model based on the predicted load demand and the photovoltaic output error of the photovoltaic system, and performing parameter updating on the initial stochastic optimization model according to the two constraint conditions to obtain a stochastic optimization model; adjusting and optimizing model constraint conditions through a fuzzy set based on Wasserstei n measurement to obtain a distribution robust opportunity constraint model, expressing the model as an SOCP model through multi-port SOP operation optimization data, and solving the model to obtain an optimization result. By integrating the energy storage system into the SOP model, greater space-time flexibility is provided for the power distribution network; during optimization, the comprehensive cost, the photovoltaic output and the load of the system are all considered, and the comprehensiveness and the accuracy of the optimization scheme are improved.

Description

Technical Field

[0001] The present invention relates to the field of distribution network optimization operation, and in particular to a distribution network SOP operation optimization method and device based on distributed blue rod. Background Art

[0002] With the popularization of carbon neutrality, distributed power sources represented by wind power and photovoltaics have been widely connected to distribution networks to achieve green development of distribution networks. However, the high penetration rate and uncertainty of distributed power sources have caused significant changes in the direction of power flow in distribution networks, from traditional one-way flow to two-way flow, and have resulted in problems such as voltage over-limit and increased peak-to-valley difference in the system, posing great challenges to the operation and control of distribution networks.

[0003] In the prior art, in response to the above problems, it is proposed to replace the traditional tie switch with an intelligent soft switch to achieve flexible interconnection between the feeders of the distribution network, making the system operation more reliable. Among them, the multi-port ESOP (MT-ESOP) is a new type of power electronic device integrating intelligent soft switches and energy storage devices. It can not only perform peak load shaving and valley filling, but also realize continuous regulation of the active power / reactive power of the distribution network, and can also be flexibly adjusted according to the real-time operating status of the distribution network.

[0004] However, the existing research on the MT-ESOP operation optimization problem still has some limitations, which can easily lead to overly expensive cost solutions, that is, the cost of MT-ESOP operation optimization is too high, and when using robust optimization methods for harsh scenarios that are not displayed, it is easy to cause the optimization solution to be too conservative or too expensive. Summary of the invention

[0005] The present invention provides a distribution network SOP operation optimization method and device based on distributed blue rod, so as to improve the comprehensive benefits of the distribution network and ensure the technical effect of distributed power supply consumption.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a distribution network SOP operation optimization method based on distributed rod, comprising the following steps:

[0007] Acquire operating parameter data of a distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain a first constraint condition and a second constraint condition;

[0008] Constructing a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and updating parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model;

[0009] The constraints of the random optimization model are converted through a preset fuzzy set based on Wasserstein metric to obtain a distributional robustness chance constraint model. The distributional robustness chance constraint model is represented as a SOCP model and solved through multi-port SOP operation optimization data to obtain an operation optimization result.

[0010] The distribution network SOP operation optimization method provided by the present invention first needs to obtain the operating parameter data of the distributed photovoltaic system, that is, obtain the current operating parameter data of the distributed photovoltaic system to be optimized, and construct the corresponding energy storage SOP model and the relaxation power flow model according to the operating parameter data, and then obtain the corresponding first constraint and second constraint by relaxing the above two models. The present invention integrates the energy storage system into the SOP model to obtain the corresponding multi-port ESOP model, that is, by storing energy and distributing it to different distribution networks, the distribution network provides greater spatiotemporal flexibility. In order to reduce the comprehensive cost of the constructed energy storage SOP model, the system also introduces the power flow calculation constraints based on the load demand forecast value and the photovoltaic output forecast value, that is, constructing the relaxation power flow model to provide corresponding constraints for the subsequent cost reduction of the optimization model.

[0011] After determining the two constraints, the system will construct an initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the target photovoltaic system, i.e., the distributed photovoltaic system. Since both the predicted load demand and photovoltaic output error will cause fluctuations in the node voltage and branch current of the distribution network, in order to further improve the accuracy of the output optimization method, the system needs to add the above two to the model parameters when constructing the stochastic optimization model. After constructing the initial stochastic optimization model, the system will adjust the constraints of the initial stochastic optimization model according to the constraints determined by the energy storage SOP model and the relaxed current model, that is, when determining the optimization plan, the comprehensive cost of the photovoltaic system, the photovoltaic output and the load are all considered as constraints to improve the comprehensiveness and accuracy of the output optimization plan.

[0012] After determining the constraints and obtaining the stochastic optimization model, the system will convert the constraints of the stochastic optimization model into a distributed robust opportunity constraint model through a fuzzy set based on the Wasserstein metric, that is, convert the stochastic optimization model into a distributed robust opportunity constraint model, so that the opportunity constraints of the node voltage and branch flow of the adjusted distribution network can satisfy all the distributed opportunity constraints in the fuzzy set. After obtaining the distributed robust opportunity constraints, the system will convert the distributed robust opportunity constraint model into a SOCP model for solution through the multi-port SOP operation optimization data, and obtain the corresponding optimization results, that is, the optimization plan for the distributed photovoltaic system.

[0013] As a preferred example, constructing the energy storage SOP model and the relaxation power flow model according to the operating parameter data specifically includes:

[0014] Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data;

[0015] The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

[0016] Since the multi-port ESOP, namely MT-ESOP, is a power electronic device composed of multiple converters sharing a DC link, in order to provide greater flexibility for the distribution network, the distribution network SOP operation optimization method provided by the present invention also integrates the energy storage system into the MT-SOP when calculating and determining the constraints of the optimization model, that is, by storing energy and reasonably distributing it to different distribution networks, the spatiotemporal flexibility of the modern distribution network is improved.

[0017] After integrating the operation data of the energy storage system and the operation data of the SOP into the energy storage SOP operation data, the corresponding energy storage SOP model can be constructed according to the energy storage SOP operation data, and the relaxation power flow model can be constructed according to the system operation data, that is, the relaxation power flow model can be constructed according to the node voltage and branch power flow data of the distributed photovoltaic system. The output data of the above two models are used as the constraints of the optimization model to provide a data basis for the subsequent construction of the stochastic optimization model.

[0018] As a preferred example, the corresponding initial stochastic optimization model is constructed based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, specifically including:

[0019] By adjusting the parameters of the predicted load demand and the photovoltaic output error of the distributed photovoltaic system through preset random variables, a random prediction error parameter is obtained;

[0020] The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

[0021] When constructing the random optimization model, the system will first adjust the parameters of the system's predicted load demand and photovoltaic output error through preset random variables, that is, the system's predicted load demand and photovoltaic output error change data are represented by random variables to avoid the complexity of the problem caused by too many variables. Therefore, each load and photovoltaic output error of the distribution network are unified into one variable, that is, the formula for the system's predicted load demand and photovoltaic output error is expressed through random variables.

[0022] After the random prediction error parameters are determined, the system can adjust the initial node voltage formula and the initial branch flow formula of the system according to the random prediction error parameters, that is, express the above two formulas through the random prediction error parameters, and construct the initial random optimization model according to the formula after parameter adjustment, providing a data basis for the subsequent construction of the distributed robust opportunity constraint model.

[0023] As a preferred example, the adjusting the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter includes:

[0024] The initial node voltage formula of the distributed photovoltaic system is parameter adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula;

[0025] The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

[0026] In order to ensure the accuracy, rationality and practicality of various constraints in the generated random optimization model, the optimization method provided by the present invention, when adjusting the parameters of the node voltage formula and the branch flow formula according to the random prediction error parameters, firstly adjusts the parameters of the above two formulas through the random prediction error parameters and the SOP output adjustment parameters, and then expresses the formulas after parameter adjustment through affine functions to obtain the corresponding node voltage parameter formula and branch flow parameter formula, thereby providing a data basis for the subsequent construction of the distributed robust opportunity constraint model.

[0027] As a preferred example, the distributed robustness opportunity constraint model is expressed as a SOCP model and solved by multi-port SOP operation optimization data to obtain an operation optimization result, including:

[0028] Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve;

[0029] Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model;

[0030] The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result.

[0031] After obtaining the distributed robust optimization model, the system will first obtain the distribution network topology data, load forecast curve and photovoltaic output forecast curve, and construct the corresponding distribution network operation model based on multi-port SOP according to the above data, and solve the operation model to obtain the corresponding multi-port SOP operation optimization data. The distributed robust opportunity constraint model is configured through the operation optimization data, that is, the multi-port SOP operation optimization problem is configured in the constructed distributed robust opportunity constraint model, so that the configured distributed robust opportunity constraint model can optimize the multi-port SOP operation problem in a targeted manner.

[0032] After completing the configuration of the split-rod opportunity constraint model, the system will determine the operation optimization problem of the multi-port SOP based on the adjusted model, express it as its corresponding SOCP model, and solve the SOCP model to obtain the corresponding solution to the operation optimization problem of the multi-port SOP, that is, the operation optimization plan of the multi-port SOP.

[0033] Accordingly, an embodiment of the present invention further provides a distribution network SOP operation optimization device based on distributed rods, the distribution network SOP operation optimization device comprising a constraint determination module, an optimization model construction module and a model conversion solution module; wherein:

[0034] The constraint condition determination module is used to obtain operating parameter data of the distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain a first constraint condition and a second constraint condition;

[0035] The optimization model building module is used to build a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and update the parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model;

[0036] The model conversion and solving module is used to convert the constraint conditions of the random optimization model through a preset fuzzy set based on Wasserstein metric, obtain a distributional robustness chance constraint model, express the distributional robustness chance constraint model as a SOCP model and solve it through multi-port SOP operation optimization data, and obtain the operation optimization result.

[0037] As a preferred example, the constraint condition determination module constructs an energy storage SOP model and a relaxation power flow model according to the operating parameter data, specifically including:

[0038] Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data;

[0039] The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

[0040] As a preferred example, the optimization model building module builds a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, specifically including:

[0041] By using a preset random variable, the prediction review demand and the photovoltaic output error of the distributed photovoltaic system are adjusted to obtain a random prediction error parameter;

[0042] The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

[0043] As a preferred example, the optimization model building module adjusts the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter, including:

[0044] According to the random prediction error and the preset SOP output adjustment parameter, the initial node voltage formula of the distributed photovoltaic system is parameter adjusted, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula;

[0045] The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

[0046] As a preferred example, the model conversion and solving module represents the distributed robust opportunity constraint model as a SOCP model through multi-port SOP operation optimization data and solves it to obtain an operation optimization result, including:

[0047] Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve;

[0048] Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model;

[0049] The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 : A flow chart of an embodiment of a distribution network SOP operation optimization method based on distributed blue rods provided by the present invention;

[0051] Figure 2 : A structural schematic diagram of an embodiment of a distribution network SOP operation optimization device based on distributed blue rods provided by the present invention;

[0052] Figure 3 : A structural diagram of the distribution network in the planning area of ​​a typical IEEE 33-node test system provided by the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Embodiment 1

[0055] Please refer to Figure 1, which is a flow chart of an embodiment of the distribution network SOP operation optimization method based on distributed blue rod provided by the present invention, including steps 101 to 103, each step is as follows:

[0056] Step 101: Acquire operating parameter data of a distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain a first constraint condition and a second constraint condition.

[0057] The distribution network SOP operation optimization method provided by the embodiment of the present invention first needs to obtain the operating parameter data of the distributed photovoltaic system, that is, obtain the current operating parameter data of the distributed photovoltaic system to be optimized, and construct the corresponding energy storage SOP model and the relaxed power flow model according to the operating parameter data, and then obtain the corresponding first constraint and second constraint by relaxing the above two models. The present invention integrates the energy storage system into the SOP model to obtain the corresponding multi-port ESOP model, that is, by storing energy and distributing it to different distribution networks, the distribution network provides greater spatiotemporal flexibility. In order to reduce the comprehensive cost of the constructed energy storage SOP model, the system also introduces the power flow calculation constraints based on the load demand forecast value and the photovoltaic output forecast value, that is, constructing the relaxed power flow model to provide corresponding constraints for the subsequent cost reduction of the optimization model.

[0058] Specifically, the energy storage SOP model and the relaxation power flow model are constructed according to the operating parameter data in this embodiment, specifically including:

[0059] Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data;

[0060] The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

[0061] Since the multi-port ESOP, namely MT-ESOP, is a power electronic device composed of multiple converters sharing a DC link, in order to provide greater flexibility for the distribution network, the distribution network SOP operation optimization method provided by the present invention also integrates the energy storage system into the MT-SOP when calculating and determining the constraints of the optimization model, that is, by storing energy and reasonably distributing it to different distribution networks, the spatiotemporal flexibility of the modern distribution network is improved.

[0062] After integrating the operation data of the energy storage system and the operation data of the SOP into the energy storage SOP operation data, the corresponding energy storage SOP model can be constructed according to the energy storage SOP operation data, and the relaxation power flow model can be constructed according to the system operation data, that is, the relaxation power flow model can be constructed according to the node voltage and branch power flow data of the distributed photovoltaic system. The output data of the above two models are used as the constraints of the optimization model to provide a data basis for the subsequent construction of the stochastic optimization model.

[0063] In this embodiment, in order to provide greater flexibility for the distribution network, the system integrates the energy storage system (ESS) into the MT-SOP. By storing energy and reasonably distributing it to different distribution networks, it is possible to provide spatiotemporal flexibility for the modern distribution network. The multi-port SOP model with energy storage constructed in this way, that is, the energy storage SOP model formula described in this embodiment is as follows:

[0064]

[0065] |P ess.t |≤P ess.max (4)

[0066] SOC (t+1) =SOC t -(P ess.t +P ess.loss.t )Δt (5)

[0067] P ess.loss.t =C ess.loss |P ess.t | (6)

[0068] SOC min ≤SOC t ≤SOC max (7)

[0069]

[0070] In the above formula, is the active power transmitted by the kth SOP at time t, is the reactive power transmitted by the kth SOP at time t, is the equipment loss of the kth SOP at time t, P ess.t is the active output of the energy storage device, S sop is the upper capacity limit of SOP, C sop.loss is the loss factor of SOP, P ess.max is the maximum output power of the energy storage device, P ess.loss.t is the equipment loss of the energy storage device at time t, SOC t is the state of charge of the energy storage device at time t, C ess.lossis the loss coefficient of the energy storage device, SOC min and SOC max are the lower and upper limits of the state of charge.

[0071] Then, constraining the above equations makes the model a non-convex quadratic constraint problem, so the formula for relaxing these constraints is as follows:

[0072]

[0073] After the constraints are relaxed, the decision variables of the MT-ESOP model, i.e., the energy storage SOP model described in this embodiment, are as follows:

[0074]

[0075] After determining the energy storage SOP model and its corresponding first constraint condition, the system will also construct the relaxed power flow model.

[0076] In order to minimize the comprehensive cost of accessing MT-ESOP, the embodiment of the present invention introduces the power flow calculation constraints based on the load demand forecast value and the photovoltaic output forecast value, and formulates the three-phase balanced power flow calculation in the distribution network with reference to the Distflow model and its convex relaxation. The relaxed SOCP formula of the Distflow model is as follows:

[0077]

[0078] In the above formula, the branch current and the square of the node voltage of the PV system are defined as additional variables is the resistance of line ij, is the impedance of line ij, is the active power transmitted by line ij, is the reactive power transmitted by line ij, is the active power demand of node j in the kth distribution network at time t, is the photovoltaic active power output of node j in the kth distribution network at time t, is the reactive power demand of node j in the kth distribution network at time t, is the photovoltaic reactive power output of node j in the kth distribution network at time t.

[0079] The decision variables of the Distflow model are as follows:

[0080]

[0081] Step 102: constructing a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and updating parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model.

[0082] After determining the two constraints, the system will construct an initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the target photovoltaic system, i.e., the distributed photovoltaic system. Since both the predicted load demand and photovoltaic output error will cause fluctuations in the node voltage and branch current of the distribution network, in order to further improve the accuracy of the output optimization method, the system needs to add the above two to the model parameters when constructing the stochastic optimization model. After constructing the initial stochastic optimization model, the system will adjust the constraints of the initial stochastic optimization model according to the constraints determined by the energy storage SOP model and the relaxed current model, that is, when determining the optimization plan, the comprehensive cost of the photovoltaic system, the photovoltaic output and the load are all considered as constraints to improve the comprehensiveness and accuracy of the output optimization plan.

[0083] Specifically, the corresponding initial stochastic optimization model is constructed based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system in this embodiment, specifically including:

[0084] By adjusting the parameters of the predicted load demand and the photovoltaic output error of the distributed photovoltaic system through preset random variables, a random prediction error parameter is obtained;

[0085] The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

[0086] When constructing the random optimization model, the system will first adjust the parameters of the system's predicted load demand and photovoltaic output error through preset random variables, that is, the system's predicted load demand and photovoltaic output error change data are represented by random variables to avoid the complexity of the problem caused by too many variables. Therefore, each load and photovoltaic output error of the distribution network are unified into one variable, that is, the formula for the system's predicted load demand and photovoltaic output error is expressed through random variables.

[0087] After the random prediction error parameters are determined, the system can adjust the initial node voltage formula and the initial branch flow formula of the system according to the random prediction error parameters, that is, express the above two formulas through the random prediction error parameters, and construct the initial random optimization model according to the formula after parameter adjustment, providing a data basis for the subsequent construction of the distributed robust opportunity constraint model.

[0088] In this embodiment, since the error in predicted load demand and photovoltaic output will cause fluctuations in node voltage and branch power flow of the distribution network, in order to avoid excessive complexity caused by too many variables, the system provided in this embodiment unifies each load and photovoltaic output error of the distribution network into one variable, namely, the random variable described in this embodiment. After adjusting the predicted load demand and photovoltaic output error according to the random variable described in this embodiment, the realized load and photovoltaic output are as follows:

[0089]

[0090] In the above formula, the random variable ( and ) belongs to the real probability distribution. In this embodiment, the variable with a superscript (~) indicates that the variable is affected by the random variable.

[0091] Furthermore, the method of adjusting the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter in this embodiment includes:

[0092] The initial node voltage formula of the distributed photovoltaic system is parameter adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula;

[0093] The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

[0094] In order to ensure the accuracy, rationality and practicality of various constraints in the generated random optimization model, the optimization method provided by the present invention, when adjusting the parameters of the node voltage formula and the branch flow formula according to the random prediction error parameters, firstly adjusts the parameters of the above two formulas through the random prediction error parameters and the SOP output adjustment parameters, and then expresses the formulas after parameter adjustment through affine functions to obtain the corresponding node voltage parameter formula and branch flow parameter formula, thereby providing a data basis for the subsequent construction of the distributed robust opportunity constraint model.

[0095] In this embodiment, after the system determines the random variables, it can calculate and determine the node voltage parameter formula and the branch power flow parameter formula according to the determined random variables. Among them, the node voltage parameter formula reflecting the impact of the prediction error and the SOP output regulation is expressed by an affine function as follows:

[0096]

[0097]

[0098] In the above formula, is the predicted voltage amplitude of the ith node of the kth distribution network, which is calculated by the independent power flow of the predicted PV output and load forecast. is the voltage change at node i caused by load forecast error, is the voltage change of node i caused by the photovoltaic output prediction error, The voltage change of node i is caused by the different SOP transmission power.

[0099] Similarly, the branch current parameter formula reflecting the impact of prediction error and SOP output regulation is expressed by affine function as follows:

[0100]

[0101] In the above formula, is the active power transmitted by the ijth branch of the kth distribution network, is the active power change of branch ij caused by load forecast error, is the active power change of branch ij caused by the PV output prediction error, It is the active power change of branch ij caused by different SOP transmission power.

[0102] In addition, the distribution network operator (DSO) in this embodiment can also obtain additional cost benefits by controlling the probability of constraint violation, and the opportunity constraints of node voltage and branch power flow reliability in the distribution network are expressed as follows:

[0103]

[0104] In the above formula, α V and α P are the node voltage and branch power flow violation risk levels, respectively, which are key parameters for directly adjusting the violation probability; V min and V max is the lower and upper limits of the node voltage; P br.max It is the upper limit of active power transmitted by the branch.

[0105] In this embodiment, since the operating goal set by the distribution network operator (DSO), that is, the system of this embodiment, is to minimize the comprehensive cost and maximize the power revenue from the upper power grid, the cost of purchasing power from the upper power grid is defined as the total net load multiplied by the electricity price, and the net load of the kth distribution network is defined as follows:

[0106]

[0107] In the above formula, and are functions of random variables, is the square of the branch current.

[0108] Considering the uncertainty of the prediction error, this embodiment provides a general chance-constrained stochastic problem model of MT-SOP with ESS, that is, the stochastic optimization model described in this embodiment, and the formula is as follows:

[0109]

[0110] st(2)-(29)(30)

[0111] Step 103: The constraints of the random optimization model are converted through a preset fuzzy set based on Wasserstein metric to obtain a distributional robustness chance constraint model, and the distributional robustness chance constraint model is represented as a SOCP model and solved through multi-port SOP operation optimization data to obtain an operation optimization result.

[0112] After determining the constraints and obtaining the stochastic optimization model, the system will convert the constraints of the stochastic optimization model into a distributed robust opportunity constraint model through a fuzzy set based on the Wasserstein metric, that is, convert the stochastic optimization model into a distributed robust opportunity constraint model, so that the opportunity constraints of the node voltage and branch flow of the adjusted distribution network can satisfy all the distributed opportunity constraints in the fuzzy set. After obtaining the distributed robust opportunity constraints, the system will convert the distributed robust opportunity constraint model into a SOCP model for solution through the multi-port SOP operation optimization data, and obtain the corresponding optimization results, that is, the optimization plan for the distributed photovoltaic system.

[0113] Specifically, in this embodiment, the distributed blue-robust opportunity constraint model is represented as a SOCP model and solved by multi-port SOP operation optimization data to obtain an operation optimization result, including:

[0114] Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve;

[0115] Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model;

[0116] The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result.

[0117] After obtaining the distributed robust optimization model, the system will first obtain the distribution network topology data, load forecast curve and photovoltaic output forecast curve, and construct the corresponding distribution network operation model based on multi-port SOP according to the above data, and solve the operation model to obtain the corresponding multi-port SOP operation optimization data. The distributed robust opportunity constraint model is configured through the operation optimization data, that is, the multi-port SOP operation optimization problem is configured in the constructed distributed robust opportunity constraint model, so that the configured distributed robust opportunity constraint model can optimize the multi-port SOP operation problem in a targeted manner.

[0118] After completing the configuration of the split-rod opportunity constraint model, the system will determine the operation optimization problem of the multi-port SOP based on the adjusted model, express it as its corresponding SOCP model, and solve the SOCP model to obtain the corresponding solution to the operation optimization problem of the multi-port SOP, that is, the operation optimization plan of the multi-port SOP.

[0119] In this embodiment, the fuzzy set is defined as a set of distributions that covers the potential distribution within a certain "distance" of the empirical distribution of finite sample data. Since the Wasserstein metric has many advantages in terms of out-of-sample performance and robustness of relying on distribution feature information, the system provided in this embodiment will use the Wasserstein metric to define the fuzzy set, that is, any two different probability distributions ( and ) and the fuzzy set in the Wasserstein sphere with radius p centered on the empirical distribution is defined in space as:

[0120]

[0121] In the above formula, is the space of Lipschitz functions for all ξ∈Ξ. is the empirical distribution of the sample Ne data. After determining the Wasserstein-based fuzzy set, the corresponding distributionally robust chance constraint model can be constructed based on the fuzzy set, that is, formula (27) and formula (28) are diffused into the distributionally robust chance constraint (DRCC) model, and the above formula (27) and formula (28) are reformulated into the form corresponding to formula (31) and formula (32), so that the adjusted formula satisfies the chance constraints of all distributions in the fuzzy set. The reformulated formula is as follows:

[0122]

[0123] Among them, the node voltage, i.e., the opportunity constraint of formula (33), can generally be expressed by separating the left and right sides of the inequality constraint as:

[0124]

[0125] The node voltage (l V.t The constraint function of (x,ξ)) is equivalent to the piecewise affine function of the decision variable and the random variable, which is as follows:

[0126]

[0127] In the above formula, the total number is twice the total number of nodes in all distribution networks connected to MT-ESOP a n.t X sop.t and b n.t ξ t They represent the voltage increments corresponding to the MT-ESOP output and the random variables (i.e., the prediction errors of load and PV output), respectively; and the two coefficients a n.t and b n.t Indicates the sensitivity of the voltage to the power in each distribution network; the last coefficient c n.t It represents the difference between the node voltage and its maximum (or minimum) threshold.

[0128] Random variable ξ t Then it contains the following random variables within the boundary (Cξ t ≤d):

[0129]

[0130] However, the DRCC problem in formula (35) has infinite dimensions due to its ambiguous probability distribution. Therefore, this embodiment approximates the DRCC problem as a CVaR problem, and the DRCC in formula (35) can be approximated as:

[0131]

[0132] By distributing α V The definition of CVaR under risk level can be expressed in formula (37) as:

[0133]

[0134] By using the duality theory, the above problem can be transformed into a tractable finite convex programming. The transformed finite convex programming formula is as follows:

[0135]

[0136] ||C T γ t.i.n -b n.t || ∞ ≤λ t (43)

[0137]

[0138] In the above formula, for any finite sample and the affine function The constraints should be satisfied, and the dual variable of this problem is defined as X dual.t =[γ t.i.n ,s t.i ,λ t ,τ t ].

[0139] Assuming that the expectation of the random variable is zero, it means that the expected value of the random component in the objective function of formula (30) can be regarded as the predicted value Therefore, this embodiment uses the power balance equation of MT-SOP in formula (1) to convert the objective function, i.e., the total net load of the distribution network in formula (30), into the deterministic part. The specific conversion formula is as follows:

[0140]

[0141] Since the first term in formula (45) is a constant given over time, the first constant can be excluded from the objective function. Therefore, the MT-ESOP operation optimization problem considering DRCC provided in this embodiment is expressed as the following SOCP model:

[0142]

[0143] st(2)-(26),(41)-(45) (46)

[0144] X t =[X n.1.t ,X n.2.t ,…,X n.nd.t ,X sop.t ,X dual.t ] T (47)

[0145]

[0146] In the above formula, the Distflow variable, MT-SOP output and dual problem variable in a day are decision variables.

[0147] In order to further illustrate the operation optimization method provided by the embodiment of the present invention, this embodiment is also described by taking a typical IEEE 33 node system as a research object. Figure 3 , Figure 3 This is a structural diagram of the distribution network in the planning area of ​​a typical IEEE 33-node test system provided by the present invention.

[0148] like Figure 3 As shown, the test system uses MT-ESOP to connect 4 distribution networks. In order to overcome the problem of low voltage in the existing test system, Figure 3 The test system shown also installs two 400kVar reactive compensators at the end of the feeder. Figure 3 The detailed location and capacity of the PV system and SOP in the distribution network of the test system shown are shown in the following table:

[0149]

[0150] Figure 3 The test system shown selected two groups of reactive compensation devices to be connected to the end of the distribution network, and the access nodes were nodes 17 and 32. For distribution network No. 1, four groups of photovoltaic devices were selected to be connected to the distribution network, and the access nodes were nodes 7, 9, 23, 28 and 32.

[0151] This embodiment is aimed at Figure 3 The test system shown sets four comparison methods, namely:

[0152] Solution 1: No MT-ESOP is connected to the distribution network.

[0153] Scheme 2: Deterministic optimization of MT-ESOP considering only the predicted PV power and load output.

[0154] Solution 3: All constraints need to be satisfied and robust optimization is performed on MT-ESOP.

[0155] Solution 4: Apply the optimization method proposed in this patent to MT-ESOP.

[0156] After determining the above four comparison methods, the test system will calculate and compare the system power loss, daily operating cost and over-limit rate of the four solutions. The results are shown in the following table:

[0157]

[0158]

[0159] When different confidence levels are set and the confidence level is the node voltage, the results of the above four solutions are shown in the following table:

[0160]

[0161] When different confidence levels are set and the confidence level is the branch flow, the results of the above four solutions are shown in the following table:

[0162]

[0163] In the above tables, the indicator "violation rate" refers to the violation rate of the reliability constraint of the power system on the bus voltage, which is defined as:

[0164]

[0165] In the above formula, Ω T represents the total time period (e.g., from 1 hour to 24 hours); the number of violations at time t represents the number of samples that violate the reliability constraint in a given sample when MT-ESOP operates at the optimal setting value at time t.

[0166] The comparison result table of four comparison schemes without given confidence, that is, the second table of this embodiment, compares the simulation results of each scenario by realizing 30 sample prediction error situations. In this table, compared with Scheme 1, Schemes 2, 3 and 4 can determine that MT-ESOP can effectively reduce network losses. Among them, Scheme 2 has the best loss reduction effect, but because it does not expect the prediction error, the over-limit rate of Scheme 2 is higher. In Scheme 3, which meets all uncertainty constraints, the violation rate of all sample cases is 0%; however, in order to ensure safety, the cost of Scheme 3 is greatly increased, and its loss reduction effect is the worst. In Scheme 4, the scheme adopts the distribution network SOP operation optimization method based on distributed robustness provided in this embodiment, so it relaxes the restrictions on the over-limit rate compared to Scheme 3, and achieves better loss reduction.

[0167] Therefore, it can be concluded from the contents of the above tables that the higher the allowable over-limit rate is, the higher the effect of reducing losses and increasing costs is. That is, by adjusting the risk level and radius of the fuzzy set, the optimization method proposed in this embodiment can achieve a balance between economy (network loss) and risk (over-limit rate).

[0168] In order to better illustrate the working principle and step flow of the distribution network SOP operation optimization method and device based on distributed rods of the present invention, you can but are not limited to refer to the relevant records above.

[0169] Accordingly, see Figure 2 , Figure 2 The schematic diagram of the structure of an embodiment of the distribution network SOP operation optimization device based on distributed rod provided by the present invention. Figure 2As shown, the distribution network SOP operation optimization device includes a constraint determination module 201, an optimization model construction module 202 and a model conversion solution module 203; wherein:

[0170] The constraint determination module 201 is used to obtain operating parameter data of the distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain the first constraint and the second constraint.

[0171] Furthermore, the constraint condition determination module 201 constructs an energy storage SOP model and a relaxation power flow model according to the operation parameter data, specifically including:

[0172] Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data;

[0173] The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

[0174] The optimization model building module 202 is used to build a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and update the parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model.

[0175] Furthermore, the optimization model building module 202 builds a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, specifically including:

[0176] By using a preset random variable, the prediction review demand and the photovoltaic output error of the distributed photovoltaic system are adjusted to obtain a random prediction error parameter;

[0177] The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

[0178] Furthermore, the optimization model building module 202 adjusts the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter, including:

[0179] According to the random prediction error and the preset SOP output adjustment parameter, the initial node voltage formula of the distributed photovoltaic system is parameter adjusted, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula;

[0180] The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

[0181] The model conversion and solving module 203 is used to convert the constraint conditions of the random optimization model through a preset fuzzy set based on the Wasserstein metric, obtain a distributional robustness chance constraint model, and represent the distributional robustness chance constraint model as a SOCP model and solve it through multi-port SOP operation optimization data to obtain an operation optimization result.

[0182] Furthermore, the model conversion and solving module 203 expresses the distributed robust opportunity constraint model as a SOCP model through multi-port SOP operation optimization data and solves it to obtain operation optimization results, including:

[0183] Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve;

[0184] Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model;

[0185] The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result.

[0186] In summary, the embodiment of the present invention provides a distribution network SOP operation optimization method and device based on distributed robustness, which constructs an energy storage SOP model and a relaxed power flow model according to the operating parameter data obtained from the distributed photovoltaic system, and obtains the first constraint and the second constraint according to the above two models; constructs an initial random optimization model based on the predicted load demand and photovoltaic output error of the photovoltaic system, and updates the parameters of the initial optimization model according to the two constraints to obtain a random optimization model; adjusts the optimization model constraints by a fuzzy set based on the Wasserstein metric to obtain a distributed robustness opportunity constraint model, and represents the model as a SOCP model and solves it through multi-port SOP operation optimization data to obtain the optimization result. By integrating the energy storage system into the SOP model, greater spatiotemporal flexibility is provided for the distribution network; the comprehensive cost, photovoltaic output and load of the system are taken into account during optimization, which improves the comprehensiveness and accuracy of the optimization scheme.

[0187] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A distribution network SOP operation optimization method based on distributed rod, characterized in that: The following steps are involved: Acquire operating parameter data of a distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain a first constraint condition and a second constraint condition; Constructing a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and updating parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model; The constraints of the random optimization model are converted through a preset fuzzy set based on Wasserstein metric to obtain a distributional robustness chance constraint model. The distributional robustness chance constraint model is represented as a SOCP model and solved through multi-port SOP operation optimization data to obtain an operation optimization result.

2. A distribution network SOP operation optimization method based on distributed blue rod as claimed in claim 1, characterized in that, The constructing of the energy storage SOP model and the relaxation power flow model according to the operating parameter data specifically includes: Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data; The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

3. A distribution network SOP operation optimization method based on distributed rod as claimed in claim 1, characterized in that, The constructing of a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system specifically includes: By adjusting the parameters of the predicted load demand and the photovoltaic output error of the distributed photovoltaic system through preset random variables, a random prediction error parameter is obtained; The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

4. A distribution network SOP operation optimization method based on distributed rod as claimed in claim 3, characterized in that, The adjusting the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter includes: The initial node voltage formula of the distributed photovoltaic system is parameter adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula; The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error parameter and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

5. A distribution network SOP operation optimization method based on distributed rod as claimed in claim 1, characterized in that, The step of expressing the distributed robust opportunity constraint model as a SOCP model and solving it through multi-port SOP operation optimization data to obtain an operation optimization result includes: Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve; Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model; The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result.

6. A distribution network SOP operation optimization device based on distributed blue rod, characterized in that: The distribution network SOP operation optimization device comprises a constraint condition determination module, an optimization model construction module and a model conversion solution module; wherein: The constraint condition determination module is used to obtain operating parameter data of the distributed photovoltaic system, and construct an energy storage SOP model and a relaxed power flow model according to the operating parameter data, and then relax the energy storage SOP model and the relaxed power flow model to obtain a first constraint condition and a second constraint condition; The optimization model building module is used to build a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, and update the parameters of the stochastic optimization model according to the first constraint condition and the second constraint condition to obtain a stochastic optimization model; The model conversion and solving module is used to convert the constraint conditions of the random optimization model through a preset fuzzy set based on Wasserstein metric, obtain a distributional robustness chance constraint model, express the distributional robustness chance constraint model as a SOCP model and solve it through multi-port SOP operation optimization data, and obtain the operation optimization result.

7. A distribution network SOP operation optimization device based on distributed blue rods as claimed in claim 6, characterized in that: The constraint condition determination module constructs an energy storage SOP model and a relaxation power flow model according to the operating parameter data, specifically including: Parsing the operating parameter data to obtain energy storage device operating data, SOP operating data and system operating power data, integrating the energy storage device operating data and the SOP operating data to obtain energy storage SOP operating data; The corresponding energy storage SOP model is constructed based on the energy storage SOP operation data, and the relaxed power flow model is constructed based on the system operation power data.

8. A distribution network SOP operation optimization device based on distributed blue rods as claimed in claim 6, characterized in that: The optimization model building module builds a corresponding initial stochastic optimization model based on the predicted load demand and photovoltaic output error of the distributed photovoltaic system, specifically including: By using a preset random variable, the prediction review demand and the photovoltaic output error of the distributed photovoltaic system are adjusted to obtain a random prediction error parameter; The initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system are adjusted according to the random prediction error parameter, and the initial random optimization model is obtained according to the adjusted node voltage parameter formula and branch power flow parameter formula.

9. A distribution network SOP operation optimization device based on distributed blue rods as claimed in claim 8, characterized in that: The optimization model building module adjusts the initial node voltage formula and the initial branch power flow formula of the distributed photovoltaic system according to the random prediction error parameter, including: According to the random prediction error and the preset SOP output adjustment parameter, the initial node voltage formula of the distributed photovoltaic system is parameter adjusted, and the adjusted formula is expressed by an affine function to obtain the node voltage parameter formula; The initial branch power flow formula of the distributed photovoltaic system is parameter-adjusted according to the random prediction error and the preset SOP output adjustment parameter, and the adjusted formula is represented by an affine function to obtain the branch power flow parameter formula.

10. A distribution network SOP operation optimization device based on distributed blue rods as claimed in claim 6, characterized in that: The model conversion and solving module expresses the distributed robust opportunity constraint model as a SOCP model through multi-port SOP operation optimization data and solves it to obtain the operation optimization result, including: Acquire distribution network topology data, load forecast curve and photovoltaic output forecast curve of the distributed photovoltaic system, and construct a distribution network operation model based on multi-port SOP according to the distribution network topology data, the load forecast curve and the photovoltaic output forecast curve; Solving the distribution network operation model to obtain the multi-port SOP operation optimization data, and configuring the distributed robustness chance constraint model according to the multi-port SOP operation optimization data to obtain the distributed robustness chance constraint optimization model; The multi-port SOP operation optimization problem is determined according to the distributed robust chance constrained optimization model, and the multi-port SOP operation optimization problem is expressed as the corresponding SOCP model, and the SOCP model is solved to obtain the operation optimization result.

Images