A fault reconfiguration method for a digital microgrid

Abstract

This invention discloses a fault reconfiguration method for digital microgrids, specifically a method for fault reconfiguration of digital microgrids considering uncertain renewable energy generation and demand-side resource aggregators. The method establishes a demand response model for load aggregators based on price elasticity coefficients and effectively characterizes the uncertainty of renewable energy generation integrated into new distribution networks through chance constraints. Furthermore, a power flow model suitable for reconfiguration of new distribution networks is established, and then a reconfiguration model for the new energy distribution system is established considering network losses and switching costs. To address the non-convexity of the reconfiguration model, the Big M method and second-order cone relaxation are employed, and intermediate variables are introduced to transform the reconfiguration model into a mixed-integer second-order cone programming problem. Simulation experiments and detailed verification results demonstrate the reliability and effectiveness of this invention in fault reconfiguration of new energy distribution systems with demand-side resources and uncertain renewable energy generation.

Description

Technical Field

[0001] This invention relates to fault reconfiguration methods, and in particular to a fault reconfiguration method for digital microgrids that considers uncertain renewable energy generation and demand-side resource aggregators, belonging to the field of power system operation and control. Background Technology

[0002] With the unprecedented increase in electricity demand, people's requirements for the quality and reliability of electricity are also rising, placing higher standards on the operation of distribution systems. Distribution system reconfiguration, as a key means of optimizing distribution network operation, can improve the reliability and economy of the entire distribution system by flexibly adjusting the network topology, and improve voltage levels, thereby enhancing power quality. However, in recent years, with the energy crisis and environmental degradation, clean energy has developed rapidly. A large number of distributed generation sources, such as photovoltaics and wind turbines, have been connected to the distribution network, and the penetration rate of new energy generation in the distribution system has been increasing year by year. The inherent uncertainty of distributed generation poses a significant challenge to distribution network reconfiguration. Furthermore, with the development of power technology, distribution systems are gradually incorporating a large number of demand-side resources. Considering demand-side resources in system reconfiguration is also of great significance for improving the economic efficiency and reliability of system operation. In conclusion, there is an urgent need to study distribution system reconfiguration methods suitable for high-proportion clean energy generation and combined with various demand-side resources. Accurately characterizing the uncertainty of distributed generation during the reconfiguration process is crucial to ensuring the effectiveness and reliability of the reconfiguration strategy.

[0003] According to relevant literature, current methods for solving distribution system reconfiguration models mainly fall into two categories: heuristic algorithms and mathematical optimization methods. Heuristic algorithms (such as genetic algorithms and particle swarm optimization) find the optimal solution through intelligent optimization. Although these algorithms can solve non-convex problems, they cannot guarantee a globally optimal solution, and their computational efficiency is low in large-scale systems, making them unsuitable for efficient utilization. Mathematical optimization algorithms, on the other hand, require transforming the non-convex model into a convex optimization problem, thereby enabling the solution to the globally optimal solution. Therefore, constructing a reasonable distribution system reconfiguration model that fully considers its integrated distributed generation and related uncertainties, comprehensively regulates demand-side resources, and transforms the distribution network reconfiguration model into a convex optimization problem to achieve optimal system reconfiguration is of great significance for promoting the efficient and optimal operation of distribution systems. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention establishes a demand response model for load aggregators and a model for energy storage aggregators. This method establishes a distributed generation output model through chance constraints to characterize its inherent uncertainty. Simultaneously, to achieve efficient solution, the implicit chance constraints are equivalently transformed into deterministic inequality constraints. Furthermore, a Distflow power flow model suitable for the reconfiguration of novel distribution systems is established, along with constraints on node voltage, branch current, network topology, and capacitor bank switching to ensure the safe and reliable operation of the distribution system. Based on this, a distribution system reconfiguration objective function is established with economic efficiency as the goal, comprehensively considering network losses and the cost of sectionalizing switching operations. To address the non-convexity of the model, a fault reconfiguration method for digital microgrids is proposed, employing the Big M method and second-order cone relaxation to transform the reconfiguration model into a mixed-integer second-order cone programming problem.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] 1. A fault reconfiguration method for a digital microgrid, comprising the following steps:

[0007] 1) A demand response model for load aggregators was established using the electricity price elasticity coefficient. Demand response measures were implemented to respond to electricity prices, thereby achieving peak shaving and valley filling of the power load and smoothing the power load curve.

[0008] 2) Model the dynamic model of power energy storage, and further reduce the operating cost of the distribution network by reasonably regulating the charging and discharging behavior of energy storage;

[0009] 3) Although renewable energy generation has low carbon and renewable characteristics, it exhibits strong uncertainty. Therefore, opportunity constraints for renewable distributed generation were established to effectively characterize the uncertainty. At the same time, assuming that its distribution follows a Gaussian distribution, it was equivalently transformed into deterministic inequality constraints. The relationship between active power and reactive power of distributed generation was established based on the power factor.

[0010] 4) Based on the characteristics of the new power distribution system reconfiguration, a suitable improved Distflow power flow constraint was established, and intermediate variables were introduced and the power flow constraint was transformed into a convex cone by using the big M method and second-order cone relaxation; node voltage, branch current, network topology and capacitor bank switching constraints were established.

[0011] 5) Taking into account network loss costs and segmented switching operation costs, an objective function for the reconfiguration model of the new energy power distribution system was established with economic efficiency as the goal. Combined with the model and constraints in steps 1) to 4), a reconfiguration model of the new energy power distribution system based on mixed integer second-order cone programming was constructed.

[0012] Preferably, step 1) specifically comprises:

[0013] The designed load aggregator model based on the electricity price elasticity coefficient utilizes a demand response strategy to achieve load smoothing, thereby reducing system network losses and improving the economics of reconfiguration. It schedules price-incentivized loads based on the electricity price elasticity coefficient to achieve peak shaving and valley filling of electricity load. It also determines the load demand after adopting time-of-use pricing based on the electricity price elasticity coefficient. The electricity load elasticity coefficient represents the percentage change in electricity demand caused by changes in electricity prices within a given time period, and its calculation method is as follows:

[0014]

[0015] Where ξ i,t ΔP represents the price elasticity coefficient of node i at time t; i,t and Δρ i,t These represent the changes in electricity demand and electricity price after node i executes its demand response at time t, respectively. and P i,t These represent the load power of node i before and after executing the demand response at time t; and ρ i,t The electricity prices at node i before and after the adoption of time-of-use pricing at time t are respectively: Represented as:

[0016]

[0017] Where ρ peak , ρ shoulder and ρ valley Electricity prices during peak, flat, and valley periods, respectively; T peak T shoulder and T valley These represent the peak and the valley periods, respectively.

[0018] To ensure the energy demand of load aggregators remains constant before and after demand response, the total power remains unchanged, as follows:

[0019]

[0020] Where T is the duration of the reconstruction optimization; Δt is the duration of a single moment.

[0021] Preferably, step 2) specifically includes:

[0022] The energy storage system configured in the system effectively improves its operational flexibility. The dynamic operation model of the energy storage is as follows:

[0023]

[0024] Q i,t The amount of electricity stored by the energy storage system configured for node i at time t; and These represent the charging power and discharging power of the energy storage system configured at node i at time t, respectively. and The charging and discharging efficiencies of the energy storage system configured for node i are respectively.

[0025] The operating power and stored power of an energy storage system must meet certain constraints to ensure its healthy operation, as expressed as:

[0026]

[0027] in The upper limit of the charge and discharge power of the energy storage system configured for node i; Q i and These represent the lower and upper limits of the amount of electricity that the energy storage system configured for node i can store, respectively.

[0028] Preferably, step 3) specifically includes:

[0029] The output power of distributed generation exhibits significant uncertainty. This method employs a chance constraint to capture this uncertainty, limiting the probability that the output power of distributed generation satisfies the constraint to a certain confidence level, expressed as:

[0030]

[0031] Where Pr(i) represents the probability of event (i) occurring; and Represent the actual and predicted power output of the distributed generation configured at node i at time t, respectively; η i The confidence level of the distributed generation output power configured for node i;

[0032] Since the implicit opportunity constraints are difficult to solve directly, it is assumed that the prediction error of distributed generation output power follows a Gaussian distribution, expressed as:

[0033]

[0034] Where σ i,t,fore The standard deviation of the output power prediction error of the distributed generation configured for node i at time t is used to obtain the following formula:

[0035]

[0036] Where Φ(i) is a random variable The cumulative distribution function of the probability distribution; according to After converting the predicted values ​​and standard deviations to a standard normal distribution, we obtain the following equation:

[0037]

[0038] Where Φ a (i) represents the cumulative distribution function of the standard normal distribution. After rearranging it, we obtain the deterministic equivalent form of the chance constraint, which is expressed as:

[0039]

[0040] Based on the operating power factor of the distributed power source, its output reactive power is expressed as:

[0041]

[0042] in δ represents the actual value of reactive power output by the distributed generation configured at node i at time t; i The power factor of distributed generation configured for node i.

[0043] Preferably, step 4) specifically includes:

[0044] By introducing the branch switch variable α ij A distflow power flow model suitable for fault reconfiguration of new energy power distribution systems was established, represented as:

[0045]

[0046]

[0047] Where α ij Let α be a binary variable representing the on / off state of branch ij. A value of 1 indicates that the branch is closed, and a value of 0 indicates that the branch is open. For faulty branches, α can be used. ij Set to 0 to avoid closing the corresponding branch in the refactoring decision; P ij,t and Q ij,t Let r represent the active power and reactive power flowing through branch ij at time t, respectively; ij and x ij These are the resistance and reactance of branch ij, respectively; I ij,t P represents the effective value of the current flowing through branch ij at time t; j,t and Q j,t These represent the injected active power and reactive power at node i, respectively. Ui represents the reactive load of node i at time t. , t represents the voltage magnitude of node i at time t; f(j) and s(j) represent the sets of parent and child nodes of node j, respectively, where the parent and child nodes represent the upstream and downstream nodes of the branch power flow in the network, respectively; Ω is the set of all branches in the distribution network;

[0048] The above Distflow power flow model has non-convex constraints and is only applicable to the case of node closure. Therefore, this method introduces intermediate variables. and By applying uncertainty constraints and relaxing them using the Big M method, the above model is extended to the set of all branches Ω; the introduced intermediate variables and uncertainty constraints are expressed as follows:

[0049]

[0050] -α ij M1≤P ij,t ≤α ij M1

[0051] -α ij M2≤Q ij,t ≤α ij M2

[0052] -α ij M3≤I ij,t ≤α ij M3

[0053] Where M1, M2, and M3 are sufficiently large positive numbers;

[0054] Combining the intermediate variables with the original Distflow model yields the transformed constraints, represented as follows:

[0055]

[0056]

[0057] The constraints obtained by further relaxation using the Big M method are expressed as follows:

[0058] mij=M4(1-αij)

[0059]

[0060] Where M4 is a sufficiently large positive number; m ij The intermediate variables required for branch ij;

[0061] To address the nonconvexity of the aforementioned power flow model, a second-order cone relaxation is applied, which is expressed as:

[0062]

[0063] During the fault reconstruction of the new energy power distribution system, the voltage of each node must also meet the upper and lower limit constraints, as shown below:

[0064]

[0065] in and These are the lower and upper limits of the voltage amplitude at node i, respectively;

[0066] If node i is a balanced node, its node voltage amplitude is always 1, which is expressed as:

[0067]

[0068] Due to limitations in line transmission capacity, branch currents must also be limited to a certain range, expressed as:

[0069]

[0070] in This represents the upper limit of the current that branch ij can transmit;

[0071] The topology of a power distribution system must satisfy connectivity and radial constraints, expressed as:

[0072]

[0073] Where f gi,t and f di,t f represents the virtual power source and virtual load of node i at time t, respectively; ij,t N represents the virtual power flow of branch ij at time t; n and N s These represent the number of nodes and the number of power sources in the power supply, respectively.

[0074] To address the undervoltage problem caused by reactive power deficit during the reconfiguration process, capacitor banks are used for reactive power compensation. The corresponding constraints are as follows:

[0075]

[0076] in This represents the reactive power compensation capacity of node i at time t; This represents the number of capacitor banks that are put into operation at node i at time t; The capacity of a single capacitor bank configured for node i; This represents the maximum number of available capacitor banks at node i.

[0077] Preferably, step 5) specifically includes:

[0078] This method aims at improving the operational economy of the power distribution system, taking into account the network loss cost and the operation cost of sectionalizing switches. The objective function of the reconstructed model is expressed as:

[0079]

[0080] Where F represents the objective function; c1 and c2 are the network loss cost coefficient and the segmented switching operation penalty cost coefficient, respectively; g 1,t g represents the network loss at time t; g2 represents the number of operations of the segmented switch; g 1,t The calculation method is expressed as follows:

[0081]

[0082] The calculation method for g2 is expressed as follows:

[0083]

[0084] Where α ij,0 This indicates the segmented switch state before reconfiguration.

[0085] The aim is to address the reconfiguration problem when demand-side resources and new energy power generation are widely integrated into new distribution systems, and to achieve optimal economic operation of the distribution system. This paper proposes a comprehensive analytical method to address the shortcomings of existing distribution system reconfiguration methods, which rarely consider the uncertainties of new energy power generation and the aggregation of demand-side resources. The designed digital microgrid fault reconfiguration method, which considers uncertain renewable energy power generation and demand-side resource aggregators, provides guidance for the operation and control of distribution systems.

[0086] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0087] (1) Because the present invention models the load aggregator as a demand response strategy based on the electricity price elasticity coefficient, it overcomes the problem that the original load curve cannot be maintained in the prior art and the operating cost of the distribution system cannot be guaranteed. Thus, it realizes the reduction of network loss cost and the improvement of system voltage quality in the reconfiguration of the distribution system by the reasonable transfer of load.

[0088] (2) Because the present invention uses opportunity constraints to deal with the uncertainty of distributed generation and uses the cumulative distribution function of uncertain variables to perform deterministic transformation of opportunity constraints, it overcomes the shortcomings of the prior art that does not consider uncertainty in the reconfiguration of the distribution system, which reduces the reliability of the reconfiguration scheme, and the shortcomings of the implicit form of opportunity constraints that are difficult to solve directly. Thus, it realizes the reconfiguration of the distribution system considering the uncertainty of distributed generation.

[0089] (3) Because the present invention establishes an improved Distflow power flow model suitable for power distribution system reconfiguration, and establishes network node voltage, branch current, network topology and capacitor bank switching constraints as well as a reconfiguration objective function with economic efficiency as the goal, the non-convexity of the model is converted into a mixed integer second-order cone programming problem. Therefore, it overcomes the shortcomings of the prior art in that it cannot achieve efficient solution of the reconfiguration model, and thus realizes the global optimal solution of the system reconfiguration model under the constraints of power distribution system operation. Attached Figure Description

[0090] Figure 1 This is a schematic diagram of the new energy power distribution system studied in this invention;

[0091] Figure 2 This is a load curve obtained by the fault reconstruction method for the new energy power distribution system designed in this invention;

[0092] Figure 3 This is the system voltage curve at 4:00 obtained by the fault reconstruction method of the new energy power distribution system designed in this invention;

[0093] Figure 4 This is the system voltage curve at 21:00 obtained by the fault reconstruction method of the new energy power distribution system designed in this invention;

[0094] Figure 5 The fault reconstruction method for new energy power distribution systems designed in this invention yields system voltage curves at 4:00 under different chance-constrained confidence levels.

[0095] Figure 6 The fault reconstruction method for new energy power distribution systems designed in this invention yields system voltage curves at 4:00 under different distributed generation penetration rates. Detailed Implementation

[0096] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0097] Example 1: Figure 1 This is a schematic diagram of the research object of the present invention. (Refer to...) Figure 1 The schematic diagram of the new energy power distribution system shown in the study illustrates a fault reconfiguration method for a digital microgrid, which includes the following steps:

[0098] 1) A demand response model for load aggregators was established using the electricity price elasticity coefficient. Demand response measures were implemented to respond to electricity prices, achieving peak shaving and valley filling of the power load and smoothing the load curve. The designed load aggregator model based on the electricity price elasticity coefficient can utilize demand response strategies to achieve load smoothing, thereby reducing system network losses and improving the economics of reconfiguration. Price-incentivized loads are scheduled based on the electricity price elasticity coefficient to achieve peak shaving and valley filling of the power load. The load demand after adopting time-of-use pricing can be determined based on the electricity price elasticity coefficient. The electricity load elasticity coefficient represents the percentage change in electricity demand caused by changes in electricity prices within a given time period, and its calculation method is as follows:

[0099]

[0100] Where ξi,t ΔP represents the price elasticity coefficient of node i at time t; i,t and Δρ i,t These represent the changes in electricity demand and electricity price after node i executes its demand response at time t, respectively. and P i,t These represent the load power of node i before and after executing the demand response at time t; and ρ i,t The electricity prices at node i before and after the adoption of time-of-use pricing at time t are respectively: It can be represented as:

[0101]

[0102] Where ρ peak , ρ shoulder and ρ valley Electricity prices during peak, flat, and valley periods, respectively; T peak T shoulder and T valley These represent the peak, the plateau, and the valley periods, respectively.

[0103] To ensure the energy demand of load aggregators remains constant before and after demand response, the total power remains unchanged, as follows:

[0104]

[0105] Where T is the duration of the reconstruction optimization; Δt is the duration of a single moment.

[0106] Reference Figure 2 The simulation experiment of this invention shows a comparison of the load curves after reconfiguration with and without the demand response strategy designed in this invention. The graph clearly shows that some load is shifted from peak hours (11:00–15:00 and 19:00–23:00) to off-peak hours (1:00–6:00 and 24:00) through demand response. Furthermore, the peak-to-valley load ratios with and without demand response are 0.0416 pu and 0.8 pu, respectively, and the operating costs of the distribution system before and after demand response are 989.2 yuan and 1025.0 yuan, respectively. These results demonstrate that considering demand response in distribution system reconfiguration can effectively smooth the load level, thereby effectively reducing the peak-to-valley difference in network load and reducing operating costs. Moreover, the distribution system only requires 6 sectionalizing switch operations in the fault reconfiguration considering demand response, while it requires 8 operations in the case without response.

[0107] Reference Figure 3 and Figure 4The simulation experiment presented by this invention compares the voltage curves of each node at 4:00 (off-peak time) and 21:00 (peak time) after reconfiguration, considering and not considering the demand response strategy designed in this invention. The comparison of the two graphs clearly shows the impact of demand response on voltage quality. More specifically, during the off-peak period (4:00), due to demand response, some peak loads are shifted to this period, which increases the total load and decreases the overall voltage level of the system. During the peak period (21:00), some loads are shifted to other periods, which decreases the total load and increases the overall voltage level of the system. These results demonstrate that considering demand response in distribution system reconfiguration can effectively reduce system operating costs by decreasing the load peak-to-valley difference, while simultaneously improving voltage quality and enhancing the economy and safety of system operation during peak hours.

[0108] 2) The energy storage system configured in the distribution network can effectively improve its operational flexibility. A dynamic model of the energy storage system is constructed, and by rationally controlling the charging and discharging behavior of the energy storage, the operating cost of the distribution network can be further reduced. The dynamic operation model of the energy storage system is as follows:

[0109]

[0110] Q i,t The amount of electricity stored by the energy storage system configured for node i at time t; and These represent the charging power and discharging power of the energy storage system configured at node i at time t, respectively. and The charging and discharging efficiencies of the energy storage system configured for node i are respectively.

[0111] The operating power and stored power of an energy storage system must meet certain constraints to ensure its healthy operation, as expressed as:

[0112]

[0113] in The upper limit of the charge and discharge power of the energy storage system configured for node i; Q i and These represent the lower and upper limits of the amount of electricity that the energy storage system configured for node i can store, respectively.

[0114] 3) Although renewable energy generation has low-carbon and renewable characteristics, it exhibits strong uncertainty. Therefore, this method designs opportunity constraints to capture the uncertainty of distributed generation. These constraints limit the probability that the output power of distributed generation satisfies the constraints to a certain confidence level, expressed as:

[0115]

[0116] Where Pr(i) represents the probability of event (i) occurring; and Represent the actual and predicted power output of the distributed generation configured at node i at time t, respectively; η i The confidence level of the distributed generation output power configured for node i.

[0117] Since the implicit chance constraints are difficult to solve directly, we can assume that the prediction error of distributed generation output power follows a Gaussian distribution, expressed as:

[0118]

[0119] Where σ i,t,fore The standard deviation of the prediction error of the output power of the distributed generation configured for node i at time t can be used to obtain the following formula:

[0120]

[0121] Where Φ(i) is a random variable The cumulative distribution function of the probability distribution. Based on... After converting the predicted values ​​and standard deviations to a standard normal distribution, we can obtain the following equation:

[0122]

[0123] Where Φ a (i) represents the cumulative distribution function of the standard normal distribution. After rearranging it, we can obtain a deterministic equivalent form of the chance constraint, expressed as:

[0124]

[0125] Based on the operating power factor of the distributed generation, its output reactive power can be expressed as:

[0126]

[0127] in δ represents the actual value of reactive power output by the distributed generation configured at node i at time t; i The power factor of distributed generation configured for node i.

[0128] Reference Figure 5The simulation experiments presented in this invention, showing the voltage curves of the entire system at 4:00 considering different confidence levels for the chance constraints, demonstrate the impact of the chance constraint confidence level on the reconfiguration of the distribution system, with the distributed generation penetration rate set at 60%. It can be seen that the overall voltage level of the system decreases as the confidence level increases. This is because at higher confidence levels, the distribution system operator's expectations for the output power of distributed generation become more conservative, leading to an increase in the net load of the entire system, which in turn reduces the system voltage level. This indicates that distribution system operators should determine a reasonable confidence level for the chance constraints based on the actual prediction accuracy of the distributed generation output power.

[0129] Reference Figure 6 The voltage curves of the entire system at 4:00, obtained from the simulation experiments of this invention considering different distributed generation penetration rates, show the impact of penetration rate on the reconfiguration of the distribution system. At this point, the distributed generator constraint is set to 0.95. It can be seen that the system voltage level increases with increasing penetration rate, because the increased distributed generation penetration rate reduces the system load level. These results indicate that an appropriate clean energy penetration rate can reduce network losses during network reconfiguration and improve voltage quality. The costs at penetration rates of 20%, 40%, 60%, and 80% are RMB 4461.5, RMB 1763.9, RMB 1194.7, and RMB 803.8, respectively. The results show that operating costs decrease with increasing distributed generation penetration rate; therefore, developing clean energy generation is of great significance for improving the quality and efficiency of the power system.

[0130] 4) Based on the characteristics of power distribution system reconfiguration, a suitable improved Distflow power flow constraint was established, and intermediate variables were introduced and the power flow constraint was transformed into a convex cone by using the big M method and second-order cone relaxation; node voltage, branch current, network topology and capacitor bank switching constraints were established.

[0131] By introducing the branch switch variable α ij A distflow power flow model suitable for fault reconfiguration of new energy power distribution systems was established, represented as:

[0132]

[0133] Where α ij Let α be a binary variable representing the on / off state of branch ij. A value of 1 indicates that the branch is closed, and a value of 0 indicates that the branch is open. For faulty branches, α can be used. ij Set to 0 to avoid closing the corresponding branch in the refactoring decision; P ij,t and Q ij,t Let r represent the active power and reactive power flowing through branch ij at time t, respectively; ij and x ijThese are the resistance and reactance of branch ij, respectively; I ij,t P represents the effective value of the current flowing through branch ij at time t; j,t and Q j,t These represent the injected active power and reactive power at node i, respectively. Ui represents the reactive load of node i at time t. , t represents the voltage magnitude of node i at time t; f(j) and s(j) represent the sets of parent and child nodes of node j, respectively, where the parent and child nodes represent the upstream and downstream nodes of the branch power flow in the network, respectively; Ω is the set of all branches in the distribution network.

[0134] The above Distflow power flow model has non-convex constraints and is only applicable to the case of node closure. Therefore, this method introduces intermediate variables. and By applying uncertainty constraints and relaxing them using the Big M method, the above model is extended to the set of all branches Ω. The introduced intermediate variables and uncertainty constraints are expressed as follows:

[0135]

[0136] -α ij M1≤P ij,t ≤α ij M1

[0137] -α ij M2≤Q ij,t ≤α ij M2

[0138] -α ij M3≤I ij,t ≤α ij M3

[0139] M1, M2, and M3 are sufficiently large positive numbers.

[0140] Combining the intermediate variables with the original Distflow model yields the transformed constraints, represented as follows:

[0141]

[0142] The constraints obtained by further relaxation using the Big M method are expressed as follows:

[0143] mij=M4(1-αij)

[0144]

[0145] Where M4 is a sufficiently large positive number; m ij These are the intermediate variables required for branch ij.

[0146] To address the nonconvexity of the aforementioned power flow model, a second-order cone relaxation is applied, which is expressed as:

[0147]

[0148] During the fault reconstruction of the new energy power distribution system, the voltage of each node must also meet the upper and lower limit constraints, as shown below:

[0149]

[0150] in and These are the lower and upper limits of the voltage amplitude at node i, respectively.

[0151] If node i is a balanced node, its node voltage amplitude is always 1, which is expressed as:

[0152]

[0153] Due to limitations in line transmission capacity, branch currents must also be limited to a certain range, expressed as:

[0154]

[0155] in This represents the upper limit of the current that branch ij can transmit.

[0156] The topology of a power distribution system must satisfy connectivity and radial constraints, expressed as:

[0157]

[0158] Where f gi,t and f di,t f represents the virtual power source and virtual load of node i at time t, respectively; ij,t N represents the virtual power flow of branch ij at time t; n and N s These represent the number of nodes and the number of power sources in the power supply, respectively.

[0159] To address the undervoltage problem caused by reactive power deficit during the reconfiguration process, capacitor banks can be used for reactive power compensation. The corresponding constraints are as follows:

[0160]

[0161] in This represents the reactive power compensation capacity of node i at time t; This represents the number of capacitor banks that are put into operation at node i at time t; The capacity of a single capacitor bank configured for node i; This represents the maximum number of capacitor banks available at node i.

[0162] 5) Taking into account network loss costs and segmented switching operation costs, an objective function for the power distribution system reconfiguration model was established with economic efficiency as the goal. Combined with the models and constraints in steps 1) to 4), a mixed integer second-order cone programming model for the power distribution system reconfiguration was established.

[0163] This method aims at improving the operational economy of the power distribution system, taking into account the network loss cost and the operation cost of sectionalizing switches. The objective function of the reconstructed model is expressed as:

[0164]

[0165] Where F represents the objective function; c1 and c2 are the network loss cost coefficient and the segmented switching operation penalty cost coefficient, respectively; g 1,t g represents the network loss at time t; g2 represents the number of operations of the segmented switch. 1,t The calculation method is expressed as follows:

[0166]

[0167] The calculation method for g2 is expressed as follows:

[0168]

[0169] Where α ij,0 This indicates the segmented switch state before reconfiguration.

Claims

1. A fault reconfiguration method for a digital microgrid, characterized in that, Follow these steps: 1) A demand response model for load aggregators was established using the electricity price elasticity coefficient. Demand response measures were implemented to respond to electricity prices, thereby achieving peak shaving and valley filling of the power load and smoothing the power load curve. 2) Model the dynamic model of power energy storage, and further reduce the operating cost of the distribution network by reasonably regulating the charging and discharging behavior of energy storage; 3) Opportunity constraints for renewable distributed generation were established to effectively characterize the uncertainty. At the same time, assuming that its distribution follows a Gaussian distribution, it was equivalently transformed into deterministic inequality constraints. The relationship between active power and reactive power of distributed generation was established based on the power factor. 4) Based on the characteristics of the new power distribution system reconfiguration, a suitable improved Distflow power flow constraint was established, and intermediate variables were introduced and the power flow constraint was transformed into a convex cone by using the big M method and second-order cone relaxation; node voltage, branch current, network topology and capacitor bank switching constraints were established. 5) Taking into account network loss costs and segmented switching operation costs, an objective function for the reconfiguration model of the new energy power distribution system was established with economic efficiency as the goal. This objective function was then combined with the models and constraints in steps 1) to 4) to construct a new energy power distribution system reconfiguration model based on mixed integer second-order cone programming. Step 3) specifically refers to: The output power of distributed generation exhibits strong uncertainty. Chance constraints capture this uncertainty, limiting the probability that the output power of distributed generation satisfies the constraints to a certain confidence level, expressed as: Where Pr(i) represents the probability of event (i) occurring; and Represent the actual and predicted power output of the distributed generation configured at node i at time t, respectively; η i The confidence level of the distributed generation output power configured for node i; Since the implicit opportunity constraints are difficult to solve directly, it is assumed that the prediction error of distributed generation output power follows a Gaussian distribution, expressed as: Where σ i,t,fore The standard deviation of the output power prediction error of the distributed generation configured for node i at time t is used to obtain the following formula: Where Φ(i) is a random variable The cumulative distribution function of the probability distribution; according to After converting the predicted values ​​and standard deviations to a standard normal distribution, we obtain the following equation: Where Φ a (i) represents the cumulative distribution function of the standard normal distribution. After rearranging it, we obtain the deterministic equivalent form of the chance constraint, which is expressed as: Based on the operating power factor of the distributed power source, its output reactive power is expressed as: in δ represents the actual value of reactive power output by the distributed generation configured at node i at time t; i The power factor of distributed generation configured for node i.

2. The fault reconfiguration method for a digital microgrid according to claim 1, characterized in that: Step 1) specifically refers to: The designed load aggregator model based on the electricity price elasticity coefficient utilizes a demand response strategy to achieve load smoothing, thereby reducing system network losses and improving the economics of reconfiguration. It schedules price-incentivized loads based on the electricity price elasticity coefficient to achieve peak shaving and valley filling of electricity load. It also determines the load demand after adopting time-of-use pricing based on the electricity price elasticity coefficient. The electricity load elasticity coefficient represents the percentage change in electricity demand caused by changes in electricity prices within a given time period, and its calculation method is as follows: Where ξ i,t ΔP represents the price elasticity coefficient of node i at time t; i,t and Δρ i,t These represent the changes in electricity demand and electricity price after node i executes its demand response at time t, respectively. and P i,t These represent the load power of node i before and after executing the demand response at time t; and ρ i,t The electricity prices at node i before and after the adoption of time-of-use pricing at time t are respectively: Represented as: Where ρ peak , ρ shoulder and ρ valley Electricity prices during peak, flat, and valley periods, respectively; T peak T shoulder and T valley These represent the peak and the valley periods, respectively. To ensure the energy demand of load aggregators remains constant before and after demand response, the total power remains unchanged, as follows: Where T is the duration of the reconstruction optimization; Δt is the duration of a single moment.

3. The fault reconfiguration method for a digital microgrid according to claim 1, characterized in that: Step 2) specifically refers to: The energy storage system configured in the system effectively improves its operational flexibility. The dynamic operation model of the energy storage is as follows: Q i,t The amount of electricity stored by the energy storage system configured for node i at time t; and These represent the charging power and discharging power of the energy storage system configured at node i at time t, respectively. and The charging and discharging efficiencies of the energy storage system configured for node i are respectively. The operating power and stored power of an energy storage system must meet certain constraints to ensure its healthy operation, as expressed as: in The upper limit of the charging and discharging power of the energy storage system configured for node i; Q i and These are the lower and upper limits of the amount of electricity that the energy storage system configured for node i can store, respectively.

4. The fault reconfiguration method for a digital microgrid according to claim 1, characterized in that: Step 4) specifically refers to: By introducing the branch switch variable α ij A distflow power flow model suitable for fault reconfiguration of new energy power distribution systems was established, represented as: Where α ij Let α be a binary variable representing the on / off state of branch ij. A value of 1 indicates that the branch is closed, and a value of 0 indicates that the branch is open. For faulty branches, α can be used. ij Set to 0 to avoid closing the corresponding branch in the refactoring decision; P ij,t and Q ij,t Let r represent the active power and reactive power flowing through branch ij at time t, respectively; ij and x ij These are the resistance and reactance of branch ij, respectively; I ij,t P represents the effective value of the current flowing through branch ij at time t; j,t and Q j,t These represent the injected active power and reactive power at node i, respectively. U represents the reactive load of node i at time t; i,t Let f(j) represent the voltage magnitude of node i at time t; f(j) and s(j) represent the sets of parent and child nodes of node j, respectively, where the parent and child nodes represent the upstream and downstream nodes of the branch power flow in the network; Ω is the set of all branches in the distribution network. The above Distflow power flow model has non-convex constraints and only applies to the case of node closure. An intermediate variable is introduced. and By applying uncertainty constraints and relaxing them using the Big M method, the above model is extended to the set of all branches Ω; the introduced intermediate variables and uncertainty constraints are expressed as follows: -a ij M1≤P ij,t ≤α ij M1 -α ij M2≤Q ij,t ≤α ij M2 -α ij M3≤I ij,t ≤α ij M3 Where M1, M2, and M3 are sufficiently large positive numbers; Combining the intermediate variables with the original Distflow model yields the transformed constraints, represented as follows: The constraints obtained by further relaxation using the Big M method are expressed as follows: m ij =M4(1-α ij ) Where M4 is a sufficiently large positive number; m ij The intermediate variables required for branch ij; To address the nonconvexity of the aforementioned power flow model, a second-order cone relaxation is applied, which is expressed as: During the fault reconstruction of the new energy power distribution system, the voltage of each node must also meet the upper and lower limit constraints, as shown below: in and These are the lower and upper limits of the voltage amplitude at node i, respectively; If node i is a balanced node, its node voltage amplitude is always 1, which is expressed as: Due to limitations in line transmission capacity, branch currents must also be limited to a certain range, expressed as: in This represents the upper limit of the current that branch ij can transmit; The topology of a power distribution system must satisfy connectivity and radial constraints, expressed as: Where f gi,t and f di,t f represents the virtual power source and virtual load of node i at time t, respectively; ij,t N represents the virtual power flow of branch ij at time t; n and N s These represent the number of nodes and the number of power sources in the power supply, respectively. To address the undervoltage problem caused by reactive power deficit during the reconfiguration process, capacitor banks are used for reactive power compensation. The corresponding constraints are as follows: in This represents the reactive power compensation capacity of node i at time t; This represents the number of capacitor banks that are put into operation at node i at time t; The capacity of a single capacitor bank configured for node i; This represents the maximum number of available capacitor banks at node i. Step 5) specifically refers to: With the goal of improving the economic efficiency of the power distribution system, and considering the network loss cost and sectionalizing switch operation cost, the objective function of the reconstructed model is expressed as: Where F represents the objective function; c1 and c2 are the network loss cost coefficient and the segmented switching operation penalty cost coefficient, respectively; g 1,t g represents the network loss at time t; g2 represents the number of operations of the segmented switch; g 1,t The calculation method is expressed as follows: The calculation method for g2 is expressed as follows: Where α ij,0 This indicates the segmented switch state before reconfiguration.

Images