Power distribution network distributed resource robust planning method and device based on cluster division

Abstract

The invention belongs to the technical field of resource planning, and particularly relates to a power distribution network distributed resource robust planning method and device based on cluster division, and the method comprises the steps: building a source-storage robust locating and sizing model as an upper model based on cluster division, and setting a target function and constraint conditions thereof; based on cluster planning, constructing an intra-cluster reliability index calculation model as a lower-layer model, and setting a target function; combining the upper-layer model and the lower-layer model to obtain a double-layer joint planning model; aiming at an island operation scene under a cluster division condition, designing a reliability merging optimization method integrated with island operation characteristics, and optimizing the double-layer joint planning model to obtain an analytical expression of a reliability index; and obtaining a joint planning scheme of the distributed resources of the power distribution network by solving the optimized double-layer joint planning model. According to the method, the source load uncertainty and the reliability index are effectively considered, and the economical efficiency, the flexibility and the reliability of power distribution network planning are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of resource planning, and particularly relates to a robust planning method and device for distributed resources in a distribution network based on cluster division. Background Art

[0002] With the rapid development of renewable energy, distributed photovoltaics (DPV) has become an important part of modern distribution networks. DPV not only helps reduce greenhouse gas emissions but also improves the economy and sustainability of distribution networks. After large-scale access of distributed power sources to the distribution network, the proportion of new energy generation in the distribution network continues to increase, changing the unidirectional power flow characteristics of traditional distribution networks and making the overvoltage problem in the distribution network more serious, affecting the safe, stable and economic operation of the distribution network. In addition, high-proportion distributed power sources are large in scale, scattered in location, numerous in quantity, and their outputs are uncertain, which greatly increases the complexity of voltage control in the distribution network. In recent years, many studies have focused on how to effectively integrate DPV into the distribution network to achieve efficient, stable and reliable power supply.

[0003] The large-scale and scattered access of distributed power sources such as wind and light to the distribution network has had an undeniable impact on the operation control and economic dispatch of the power system, making the traditional centralized control mode face problems such as complex control and "curse of dimensionality". Therefore, the optimized development of distribution network zoning is particularly important. When large-scale distributed photovoltaics are connected to the distribution network, due to the wide range of installation points of distributed photovoltaics, the variable dimension in the centralized planning model increases sharply, and the model will become too complex and difficult to solve.

[0004] Currently, the calculation of reliability indexes for distribution networks containing distributed photovoltaics is mainly divided into simulation methods and analytical methods. However, most of these two methods are based on an unchanging network structure and DG access. Before the completion of the planning, both the distribution network grid structure and DG (distributed power source) access are uncertain. There is currently a distribution network grid planning method based on power consumption reliability. This method divides the distribution network planning process into two layers. The upper layer is the distribution network planning, and the lower layer is the calculation of reliability indexes. The distribution network planning results obtained from the upper layer are transmitted to the lower layer, and the reliability index results calculated by the lower layer are returned to the upper layer for iterative solution between the upper and lower layers. However, there are still some problems with existing methods: 1) Currently, a centralized planning model for the distribution network is established on the basis of considering the investment and operation costs of distributed photovoltaics, which is more applicable when the installation ratio of distributed photovoltaics is low. When large-scale distributed photovoltaics are connected to the distribution network, the variable dimension in the centralized planning model increases sharply, and the model will become too complex and difficult to solve.

[0005] 2) The output of distributed renewable energy is volatile, and the demand of the load is also uncertain. If the distribution network is planned under deterministic scenarios, when the output of distributed renewable energy is the largest and the load demand is the smallest, it may lead to reverse power flow and even the collapse of the distribution network. Therefore, it is extremely important to comprehensively consider the impact of source-load uncertainty and ensure the safe and stable operation of the distribution network under the worst conditions.

[0006] 3) For the distribution network before planning, both distributed renewable energy and the grid structure are uncertain. Considering the volatility of the power output of distributed renewable energy and the power demand of the load, it is necessary to further study the calculation of reliability indicators to ensure the effectiveness of the distribution network planning scheme while considering reliability indicators. Summary of the Invention

[0007] Based on the above deficiencies in the prior art, the object of the present invention is to provide a robust planning method and device for distributed resources in a distribution network based on cluster division. Through a two-layer joint planning model based on cluster division, it effectively takes into account source-load uncertainty and reliability indicators, significantly improves the economy, flexibility and reliability of distribution network planning, and at the same time optimizes the calculation efficiency and distributed energy consumption capacity.

[0008] To achieve the above object, the present invention provides a robust planning method for distributed resources in a distribution network based on cluster division, including the following steps: S1. Based on cluster division, aiming at the source-load uncertainty within the cluster, introduce an uncertainty adjustment parameter to characterize the range of source-load uncertainty, establish a source-storage robust site-selection and sizing model as the upper-layer model, and set the objective function and its constraint conditions of the upper-layer model; S2. Based on cluster planning, introduce a reliability indicator, construct a reliability indicator calculation model within the cluster as the lower-layer model, and set the objective function of the lower-layer model; S3. Combine the upper-layer model and the lower-layer model to obtain a two-layer joint planning model; S4. For the island operation scenario under the condition of cluster division, design a reliability merging and optimization method integrating island operation characteristics, optimize the two-layer joint planning model, and obtain an analytical expression of the reliability indicator; S5. By solving the optimized two-layer joint planning model, obtain a joint planning scheme for distributed resources in the distribution network.

[0009] As a preferred solution of the present invention, in S1, the objective function of the upper-layer model includes the annual investment and operation and maintenance costs of photovoltaic, the annual income of photovoltaic, and the annual investment cost of energy storage. For any cluster k divided, aiming at site-selection and sizing within the cluster, establish the objective function F of the upper-layer model as: ; In the formula, F 1 is the annual investment cost of photovoltaic and energy storage in cluster k; F 2 is the annual operation and maintenance cost of photovoltaic and energy storage in cluster k; F 3 is the annual income of photovoltaic in cluster k.

[0010] As a preferred solution of the present invention, F 1 is expressed as: ; In the formula, j is the node index in cluster k, is the set of nodes included in cluster k, is the number of nodes included in cluster k; is the investment cost per megawatt of photovoltaic; is the photovoltaic planned capacity of node j in cluster k; is the discount rate; is the discounted years of photovoltaic; is the investment cost per megawatt of energy storage; is the energy storage planned capacity of node j in cluster k; r is the discounted years of energy storage; F 2 is expressed as: ; In the formula, is the annual fixed maintenance cost per megawatt of photovoltaic; is the annual fixed maintenance cost per megawatt of energy storage; F 3 is expressed as: ; In the formula, t represents the moment, and T is the number of moments; is the on-grid electricity price of distributed power sources; is the subsidy electricity price of distributed power sources; is the output limit per megawatt of photovoltaic under given sunshine conditions at moment t; is the active power load of node j at moment t.

[0011] As a preferred solution of the present invention, the constraint conditions of the upper-layer model include: Energy storage constraint: ; In the formula, , are the state of charge of the energy storage of node j at moment t and moment t-1, respectively; , are the minimum and maximum state of charge of the energy storage, respectively; is the charge and discharge efficiency of the energy storage; is the energy storage charging power of node j at time t; is the energy storage discharging power of node j at time t; is the charging 01 variable; is the discharging 01 variable; when it means the energy storage device is in the charging state at time t, and at this time is 0 to ensure that the energy storage device cannot charge and discharge simultaneously; similarly, when it means the energy storage device is in the discharging state at time t, and at this time must be 0; if and it means the energy storage device is not charging or discharging at time t; 、 are the maximum power limits for energy storage charging and discharging respectively; Power balance constraint: ; ; In the formula, is the set of upstream nodes of node j, and i represents one of the upstream nodes; is the active power flowing through the line between node i and node j at time t; is the resistance of the line between node i and node j; is the square value of the current of the line between node i and node j; is the actual active power generation of the PV at node j at time t; is the total active power transmission from node j to its downstream nodes at time t; is the set of downstream nodes of node j, and c represents one of the downstream nodes; is the reactive power flowing through the line between node i and node j at time t; is the reactance of the line between node i and j; is the reactive power load of node j at time t; is the actual reactive power generation of the PV at node j at time t; is the total reactive power transmission from node j to its downstream nodes at time t; PV capacity constraint allowed for each node in the cluster: ; In the formula, is the PV capacity of node j; is the maximum value of the load demand of node j in the cluster; Power flow constraint: ; ; ; Wherein, is the square of the voltage value of node i at time t; , are the upper and lower limits of the voltage square respectively; is the maximum current limit; Uncertainty constraint: ; Wherein, U represents a box-type uncertainty set, u is a specific variable vector in the uncertainty set, that is ; , , At this time, they respectively represent the uncertain variables of the active load power, reactive load power, and photovoltaic output of node j in cluster k; , , are the predicted values of the active load power, reactive load power, and photovoltaic output of node j in cluster k respectively; , , are the active load fluctuation amplitude, reactive load fluctuation amplitude, and photovoltaic output fluctuation amplitude respectively; , , are the uncertainty adjustment parameters introduced for the active load power, reactive load power, and photovoltaic output respectively, and the values are all between [-1, 1], indicating the proportion of the active load power, reactive load power, and photovoltaic output fluctuating up and down within the fluctuation range during the simulation period; , , are the fluctuation deviations of the active load power, reactive load power, and photovoltaic output of node j in cluster k respectively.

[0012] As a preferred solution of the present invention, in S2, the objective function of the lower-layer model is: ; Wherein, F 4 is the power outage loss cost of users in cluster k; is the average load value of node j; is the failure rate of the jth node; is the average power outage time of the jth node in case of failure; is the loss cost caused when node j is powered off based on the average load of node j; and are reliability indicators; Wherein: ; ; In the formula, is the total number of node connection branches in cluster k; is the probability of the cluster operating in island mode; is the maximum interruption amount of the interruptible average node load for node j; is the maximum interruption time of the interruptible average node load for node j.

[0013] As a preferred solution of the present invention, in S4, the reliability merging and optimization method integrating island operation characteristics is as follows: by the reverse flow calculation of reliability indexes, the reliability indexes of each component in each branch network are equivalent to the branch nodes; the branch network is equivalently simplified into components through the reverse flow merging method, and the network becomes a simplified network after the equivalent of the branch network; the influence of component failures in the main network on the branch nodes is quickly evaluated for this simplified network, and an analytical expression of the reliability indexes of the branch nodes is obtained.

[0014] As a preferred solution of the present invention, analytical expressions of reliability indexes are designed respectively for case 1 where the cluster tie switch is a circuit breaker and case 2 where the cluster tie switch is a disconnecting switch.

[0015] As a preferred solution of the present invention, the specific design method of the analytical expression of the reliability index is as follows: Case 1: When the cluster tie switch is a circuit breaker, the failure rate and the mean time to failure of the j-th node are as follows: ; ; In the formula, is the power outage rate obtained by forward flow merging of the branch node q as a main network node; is the power outage rate obtained by reverse flow transfer of the branch node q as a main network node; is the power outage rate transferred from node c to the branch node q by reverse flow; is the mean time to failure obtained by forward flow merging of the branch node q as a main network node; is the mean time to failure obtained by reverse flow transfer of the branch node q as a main network node; is the mean time to failure transferred from node c to the branch node q by reverse flow; Case 2: When the cluster tie switch is a disconnecting switch, the failure rate and the mean time to failure of the j-th node are as follows: ; ; Finally, a mathematical expression related to the failure rate of the load point, the duration of each failure, and the probability variable of the cluster island operation is obtained, which is and The analytical expression of: ; ; In the formula, is the failure rate of the j-th node obtained by the reverse transfer method; is the average power outage time of the failure of the j-th node obtained by the reverse transfer method.

[0016] As a preferred solution of the present invention, when a failure occurs in the cluster, if the conditions for island operation cannot be met, it is necessary to consider the maximum interruption time and the maximum interruption load of the interruptible load, and set the following constraint conditions: Maximum interruption time constraint: ; In the formula, is the upper limit of the maximum interruption time of the interruptible average node load for node j; Maximum interruption amount constraint: ; In the formula, is the upper limit of the maximum interruption amount of the interruptible average node load for node j; In addition, power balance constraints, power flow constraints, and uncertainty constraints are set, and the setting methods are the same as the relevant constraint conditions in the upper layer model.

[0017] A distributed resource robust planning device for a distribution network based on cluster division includes a memory, a processor, and a computer program stored on the memory and capable of running on the processor. The above method is implemented by the processor executing the computer program.

[0018] The beneficial effects of the present invention are: By establishing a two-layer joint planning model, the present invention effectively solves the problem that it is difficult to take into account the source-load uncertainty and reliability indicators in traditional planning. In the upper layer of the model, a source-storage robust site selection and sizing model is established for the source-load uncertainty within the cluster, and an uncertainty adjustment parameter is introduced to flexibly cope with the volatility in different scenarios; in the lower layer, considering the characteristics of cluster island operation, an analytical expression of the reliability index is proposed to optimize the calculation of the reliability index within the cluster. In addition, the present invention shows significant advantages in terms of calculation efficiency and planning cost. The calculation time is greatly shortened compared with centralized planning, and it is superior to traditional methods in terms of the annual investment cost of photovoltaic and energy storage, the annual revenue of photovoltaic, network loss, operation and maintenance cost, and the cost of user power outage loss.

[0019] Through simulation verification, the present invention can flexibly adjust the conservativeness of the planning scheme under different uncertainty adjustment parameters, ensuring the safe and stable operation of the distribution network under complex conditions. Compared with the centralized planning, it not only increases the PV access ratio, optimizes the energy storage configuration, but also significantly improves the reliability index, reducing the number and duration of power outages. In addition, the planning method based on cluster division performs excellently in terms of economy. While reducing the total cost, it further enhances the operation efficiency and economy of the distribution network, providing strong support for the efficient integration of distributed energy and the sustainable development of the distribution network.

[0020] In summary, based on considering the uncertainties of power sources and loads and the reliability index, the present invention realizes the optimization of the efficiency, economy, and reliability of the distribution network planning, has significant technical advantages and practical application value, and provides an innovative and effective solution for the distributed resource planning of modern distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is the process schematic diagram of the present invention; Figure 2 is the distribution network topology diagram during the verification process of the present invention; Figure 3 is the comparison diagram of the PV planned capacity under different parameter schemes during the verification process of the present invention; Figure 4 is the comparison diagram of the energy storage planned capacity under different parameter schemes during the verification process of the present invention; Figure 5 is the schematic diagram of the average number of power outages under different parameter schemes during the verification process of the present invention; Figure 6 is the schematic diagram of the average power outage duration under different parameter schemes during the verification process of the present invention; Figure 7 is the comparison diagram of the PV planned capacity under different planning schemes during the verification process of the present invention; Figure 8 is the comparison diagram of the energy storage planned capacity under different planning schemes during the verification process of the present invention; Figure 9 is the schematic diagram of the average number of power outages under different planning schemes during the verification process of the present invention; Figure 10 is the schematic diagram of the average power outage duration under different planning schemes during the verification process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The following further describes the embodiments of the present invention with reference to the drawings: Embodiment 1 As Figure 1 shown, the distributed resource robust planning method for the distribution network based on cluster division includes the following steps: S1. Based on cluster division, aiming at the uncertainty of power generation and load within the cluster, introduce an uncertainty adjustment parameter to characterize the range of power generation and load uncertainty, establish a robust siting and sizing model of power source - storage as the upper - layer model, and set the objective function and its constraint conditions of the upper - layer model; S2. Based on cluster planning, introduce a reliability index, construct a calculation model of the reliability index within the cluster as the lower - layer model, and set the objective function of the lower - layer model; S3. Combine the upper - layer model and the lower - layer model to obtain a two - layer joint planning model; S4. For the island operation scenario under the condition of cluster division, design a reliability merging and optimization method integrating the characteristics of island operation, optimize the two - layer joint planning model, and obtain an analytical expression of the reliability index; S5. By solving the optimized two - layer joint planning model, obtain the joint planning scheme of distributed resources in the distribution network.

[0023] For the island operation scenario under the condition of cluster division, a robust joint planning method of distributed resources based on cluster division is proposed, and reliability optimization and merging are carried out in combination with the characteristics of island operation. Various well - known methods can be used for cluster division.

[0024] In S1, the objective function of the upper - layer model includes the annual investment and operation and maintenance costs of photovoltaic, the annual income of photovoltaic, and the annual investment cost of energy storage. For any cluster k divided (aiming to identify the "worst" scenario of uncertain variables and minimize the cost within the cluster based on the worst scenario), taking the siting and sizing within the cluster as the goal, establish the objective function F of the upper - layer model as: ; In the formula, F 1 is the annual investment cost of photovoltaic and energy storage within cluster k; F 2 is the annual operation and maintenance cost of photovoltaic and energy storage within cluster k; F 3 is the annual income of photovoltaic within cluster k.

[0025] F 1 is expressed as: ; In the formula, j is the node index within cluster k, is the set of nodes included in cluster k, is the number of nodes included in cluster k; is the investment cost per megawatt of photovoltaic; is the planned capacity of photovoltaic at node j within cluster k; is the discount rate; is the discounted number of years of photovoltaic; is the investment cost per megawatt of energy storage; is the planned energy storage capacity of node j in cluster k; r is the annualized number of years of energy storage; F 2 is expressed as: ; In the formula, is the annual fixed maintenance cost per megawatt of photovoltaic; is the annual fixed maintenance cost per megawatt of energy storage; F 3 is expressed as: ; In the formula, t represents the time, and T is the number of time instances; is the on-grid electricity price of distributed power generation; is the subsidy electricity price of distributed power generation; is the output limit per megawatt of photovoltaic under given sunshine conditions at time t; is the active power load of node j at time t.

[0026] In this embodiment, when calculating each formula, according to the need, dimension removal or normalization processing can be performed.

[0027] The constraint conditions of the upper-layer model include: Energy storage constraint: ; In the formula, , are the state of charge of the energy storage of node j at time t and time t-1, respectively; , are the minimum and maximum state of charge of the energy storage, respectively; is the charge-discharge efficiency of the energy storage; is the charging power of the energy storage of node j at time t; is the discharging power of the energy storage of node j at time t; is the charging 01 variable; is the discharging 01 variable; when , it means that the energy storage device is in the charging state at time t, and at this time is 0 to ensure that the energy storage device cannot charge and discharge simultaneously; similarly, when , it means that the energy storage device is in the discharging state at time t, and at this time must be 0; if and , it means that the energy storage device is not charging or discharging at time t; , are the maximum power limits of energy storage charging and discharging, respectively; Power balance constraint: ; ; In the formula, is the set of upstream nodes of node j, and i represents one of the upstream nodes; is the active power flowing through the line between nodes i and j at time t; is the resistance of the line between nodes i and j; is the square value of the current in the line between nodes i and j; is the actual active power generation of the PV at node j at time t; is the total active power transmission from node j to its downstream nodes at time t; is the set of downstream nodes of node j, and c represents one of the downstream nodes; is the reactive power flowing through the line between nodes i and j at time t; is the reactance of the line between nodes i and j; is the reactive power load of node j at time t; is the actual reactive power generation of the PV at node j at time t; is the total reactive power transmission from node j to its downstream nodes at time t; PV capacity constraint for each node in the cluster: ; In the formula, is the PV capacity of node j; is the maximum value of the load demand of node j in the cluster; Power flow constraint: ; ; ; In the formula, is the square of the voltage value of node i at time t; , are the upper and lower limits of the square of the voltage respectively; is the maximum current limit; Uncertainty constraint (i.e., introducing uncertainty adjustment parameters to describe the uncertainty range of source and load): ; In the formula, U represents the box-type uncertainty set, u is the specific variable vector in the uncertainty set, that is ; , , At this time, they respectively represent the uncertain variables of the active load power, reactive load power, and PV output of node j in cluster k (the variables represented by these symbols are "uncertain variables", that is, these variables are not fixed values in the current model, but random variables that change with uncertainty, without changing the original definition, and the superscript T represents transpose); and and are respectively the predicted values of the active load power, reactive load power, and PV output of node j in cluster k; and and are respectively the fluctuation ranges of the active load, reactive load, and PV output; and and are respectively the introduced uncertain adjustment parameters for the active load power, reactive load power, and PV output, and their values are all between [-1, 1], indicating the proportion of the active load power, reactive load power, and PV output fluctuating up and down within the fluctuation range during the simulation period; and and are respectively the fluctuation deviations of the active load power, reactive load power, and PV output of node j in cluster k.

[0028] For the uncertain constraints, the purpose of this formula is to handle and consider the impact of uncertainty factors on the system performance in the optimization problem of the power system, especially in terms of planning cost and system constraints. By considering the uncertain factors, the system planning can be made more robust and can effectively adapt to different changing conditions.

[0029] The formula describes the fluctuation ranges of the active power, reactive power, and PV output in the power system under uncertain conditions. Each formula adopts the form of a box-type uncertainty set to represent the variation of the corresponding power within the uncertainty fluctuation range. Through the fluctuation range and and and the fluctuation direction and and , the uncertainty performance of the power can be flexibly controlled. The fluctuation range represents the magnitude of the power change and describes the intensity of the power fluctuation in the system. Larger and and values represent larger power fluctuations in the system, and smaller and and values indicate smaller fluctuations. The fluctuation direction and and It reflects the directionality of fluctuations, which may be upward or downward fluctuations. By adjusting 、 、 , it is possible to control whether the power fluctuations have a positive or negative impact on the system. Decision-makers can independently select 、 、 within the value range to flexibly adjust the direction of power fluctuations. If 、 、 are closer to 1 or -1, it means that the power is at the maximum positive or negative fluctuation, and the system's response to uncertainty will be more conservative; conversely, if 、 、 are close to 0, it means that the power fluctuations are small and the solution is more aggressive.

[0030] By introducing 、 、 and 、 、 , the uncertainty performance of power can be precisely controlled. This enables the system to make dynamic adjustments according to the actual situation and has high adaptability. Considering the power fluctuations, especially for the fluctuations of load and renewable energy, it can make the system more robust. By appropriately adjusting the fluctuation amplitude and direction, the planning solution can be made more reliable in the face of uncertainty and reduce the sensitivity to future uncertainty changes. Using the fluctuation amplitude 、 、 and the fluctuation direction 、 、 as parameters can more accurately describe the fluctuation behavior of power, so that the optimization model is more precise in dealing with uncertainty. This helps to reduce resource waste in actual operation and improve the overall efficiency of the system.

[0031] In S2, the objective function of the lower-level model is: ; In the formula, F 4 is the power outage loss cost of users within cluster k; is the average load value of node j; is the failure rate of the j-th node (usually the number of power outages divided by the time period is the failure rate, so the average number of power outages can be obtained through the failure rate); is the average power outage time of the j-th node's failure; is the average load based on node j, and is the loss cost caused by the power outage of node j; and is the reliability index; This formula takes into account the reliability index generated by the island operation of the cluster after a failure occurs within the cluster. If the conditions for island operation cannot be met, the maximum interruption time and the maximum interruption load of the interruptible load are considered.

[0032] F 4 In: ; ; In the formula, is the total number of branch circuits connecting nodes in cluster k; is the probability of island operation of the cluster; is the maximum interruption amount of the interruptible average node load for node j; is the maximum interruption time of the interruptible average node load for node j.

[0033] In S4, the reliability merging and optimization method integrating island operation characteristics is as follows: through the reverse calculation of reliability indexes, the reliability indexes of each component in each branch network are equivalent to the branch nodes; the branch network is equivalently simplified into components through the reverse merging method, and the network becomes a simplified network after the equivalence of the branch network; the impact of component failures in the main network on the branch nodes is quickly evaluated for this simplified network, and an analytical expression of the reliability indexes of the branch nodes is obtained.

[0034] Analytical expressions of reliability indexes are designed respectively for Case 1 where the cluster tie switch is a circuit breaker and Case 2 where the cluster tie switch is a disconnecting switch.

[0035] The specific design method of the analytical expression of the reliability index is as follows: Case 1: When the cluster tie switch is a circuit breaker, the failure rate and the mean time to failure of the j-th node based on forward merging are as follows: ; ; In the formula, is the outage rate obtained by forward merging with the branch node q as a main network node; is the outage rate obtained by reverse transfer with the branch node q as a main network node; is the outage rate transferred from node c to branch node q in reverse; is the mean time to failure obtained by forward merging with the branch node q as a main network node; is the mean time to failure obtained by reverse transfer with the branch node q as a main network node; is the average power outage time transferred counter - currently from node c to branch connection point q; Both forward merging and counter - current transfer are well - known techniques. Counter - current transfer is mainly used to consider how failures in downstream components (such as nodes in a branch network, like node j) affect the reliability of upstream nodes (such as branch connection point q) through the network. At the same time, the upstream reliability also affects the reliability of downstream nodes through forward merging. In a distribution network, the reliability index of branch connection point q is transferred to node j through forward merging. Therefore, in order to calculate the reliability of node j, the reliability of branch connection point q must be calculated first. Finally, through counter - current and forward - current calculation methods, an accurate reliability assessment is obtained by integrating the counter - current and forward - current reliability indices of each node, that is, the reliability indices of each node j in the branch network are obtained.

[0036] In the part remaining after removing the main network from a distribution network, any connected network is called a branch network. The connection point between the branch network and the main network is called a branch connection point, and node j represents any node in the distribution network. First, the counter - current transfer of reliability parameters starts from the load point at the end of the branch network until the branch connection point. Subsequently, all branch networks are equivalent to components, and the forward - current reliability index of the branch connection point is calculated through failure mode effect analysis (FMEA). Then, the forward - current merging calculation of the reliability index of the branch network is carried out. The reliability index of each node in the branch network is obtained by adding the counter - current transfer index and the forward - current merging index.

[0037] Case 2: If the cluster tie - switch is a disconnecting switch, then the failure rate of the j - th node and the mean time to failure are as follows: ; ; Finally, a mathematical expression related to the failure rate of the load point, the duration of each failure, and the cluster islanding operation probability variable is obtained, which is the and analytical expressions: ; ; In the formula, is the failure rate of the j - th node obtained by the counter - current transfer method; is the mean time to failure of the j - th node obtained by the counter - current transfer method.

[0038] and The analytical expression first calculates the reliability index in a reverse flow manner, and the reliability indices of the components in each branch network are equivalent to the branch nodes. The branch network is equivalently simplified into components through the reverse flow merging method. After the equivalent simplification of the branch network, the network becomes a simplified network. The simplified network can quickly evaluate the impact of component failures in the main network on the branch nodes, and obtain the expression of the reliability index of the branch nodes. Different clusters contain each branch network, and their distribution is relatively reasonable. When distributed renewable energy is connected to the nodes in the cluster, the cluster has a certain probability of achieving island operation. Therefore, in the calculation of the reliability index, the probability of island operation of the cluster needs to be considered. When a component outside the main network fails, the island operation can be achieved by instructing the control of the tie switch (such as a disconnecting switch or a circuit breaker). Due to the parameter differences between the disconnecting switch and the circuit breaker, the reliability index calculation needs to be analyzed separately for the two cases, and finally the reliability index considering the island operation of the cluster is obtained.

[0039] When a fault occurs within the cluster, if the conditions for island operation cannot be met, the maximum interruption time and the maximum interruption load of the interruptible load need to be considered, and the following constraints are set: Maximum interruption time constraint: ; In the formula, is the upper limit of the maximum interruption time of the interruptible average node load for node j; Maximum interruption quantity constraint: ; In the formula, is the upper limit of the maximum interruption quantity of the interruptible average node load for node j; In addition, power balance constraints, power flow constraints, and uncertainty constraints are set, and the setting methods are the same as the relevant constraint conditions in the upper-layer model.

[0040] The verification process is as follows: To verify the practicality of the above theory, a simulation verification was carried out using a certain actual distribution network as a case. The total load access capacity of the distribution network is 22.8 MVA, and the total access capacity of the photovoltaic power source is 7.5 MW. The geographical layout of the distribution network and the cluster distribution are as Figure 2 shown, and the cluster division results are shown in Table 1.

[0041] Table 1 Cluster division results

[0042] To verify the flexible adjustable ability of the uncertainty adjustment parameters proposed in this embodiment in the conservativeness of the adjustment scheme, 3 groups of uncertainty adjustment parameters are selected and compared in terms of energy storage access capacity, photovoltaic access capacity, and reliability index. The active load power and reactive load power uncertainty adjustment parameters in the 3 groups of models , , The selection results are as follows: Scenario 1: Select . In this case, the active load power, reactive load power, and PV output are all equal to the predicted values.

[0043] Scenario 2: During the planning process, the active load power and reactive load power at 30 nodes reach the minimum value of the prediction interval, that is , and the PV output at 10 nodes reaches the maximum value of the prediction interval .

[0044] Scenario 3: Select the worst-case scenario among all robust planning scenarios. The active load power and reactive load power at 75 nodes reach the minimum value of the prediction interval , and the PV output at 20 nodes reaches the maximum value of the prediction interval .

[0045] The PV planning capacity and energy storage access capacity under the comparison of the three groups of scenarios are as shown in Figure 3 , Figure 4 . The results of the average number of power outages and average power outage time are as shown in Figure 5 , Figure 6 .

[0046] According to Figure 3 shown, as the uncertainty adjustment parameter increases, the PV planning capacity in the planning results shows a gradually decreasing trend. This is mainly because the conditions faced are becoming increasingly severe. To ensure the reliability of power supply and the economy of the planning scheme and maintain the normal operation of the cluster, it is necessary to correspondingly reduce the PV access volume and ensure that the PV penetration rate of the cluster remains at a low level.

[0047] Further observation Figure 4 reveals that as the conservativeness of the planning scheme increases, the energy storage access capacity also gradually increases. This is because compared with Scenario 1, the active power load and reactive power load in Scenarios 2 and 3 are smaller, while the PV output is larger. Therefore, more energy storage systems are needed to balance the uncertainty of the PV output.

[0048] From Figure 5 , Figure 6It is concluded that as the uncertainty adjustment parameter increases, the average number of power outages and the average power outage time of nodes under different schemes also increase. As the uncertainty increases, the obtained schemes become more and more conservative, the reliability indexes of nodes all decrease, and the average power outage time and the average number of power outages of nodes all increase. This is because the divided clusters are changing and the power supply durations are different, resulting in a large degree of change in the reliability indexes. When the uncertainty increases, in order to ensure the safe, stable and economic operation of the clusters, it is necessary to increase the conservatism of the schemes.

[0049] The specific costs are shown in Table 2, where A represents the annual investment cost of photovoltaic and energy storage (10,000 yuan), B represents the annual revenue cost of photovoltaic (10,000 yuan), C represents the annual operation and maintenance cost of photovoltaic and energy storage (10,000 yuan), D represents the annual network loss cost of the cluster (10,000 yuan), E represents the user power outage loss cost (10,000 yuan), and G represents the total cost (10,000 yuan).

[0050] Table 2 Planning costs in different schemes

[0051] By deeply analyzing the data in Table 2, it can be observed that as the adjustment of the uncertain parameter gradually increases, although the annual investment and operation and maintenance costs of photovoltaic decrease, the annual revenue of photovoltaic shows a downward trend. At the same time, the network loss, the annual operation and maintenance cost of energy storage, the user power outage loss cost, etc. also increase accordingly. Behind this phenomenon, mainly because when more fully considering the uncertainties faced by the distribution network in the planning process, the obtained planning schemes tend to be more conservative. Although this conservatism helps to cope with potential risks, it will also bring an increase in power outage losses, resulting in an increase in the total cost.

[0052] To further illustrate the superiority of cluster division for distribution network planning, the distribution network planning method based on cluster division proposed in this embodiment is compared with the centralized planning method that does not divide clusters and directly performs overall calculations. The minimum values of the prediction intervals are taken for the active load power and reactive load power of 30 nodes, and the maximum values of the prediction intervals are taken for the photovoltaic output of 10 nodes. The comparison of the calculation time and the total planning cost under the two methods is shown in Table 3 and Table 4 respectively, and the photovoltaic access capacity, energy storage access capacity and user reliability indexes are as Figure 7 , Figure 8 , Figure 9 , Figure 10 .

[0053] From Figure 7 and Figure 8It can be concluded that the photovoltaic planning capacity based on cluster division planning is larger than that of centralized planning, and the total energy storage access capacity obtained by cluster division planning is smaller than that of centralized planning, indicating that the planning scheme based on cluster division can help increase the proportion of photovoltaic access and complete more distributed power consumption.

[0054] Depend on Figure 9 , Figure 10 It is concluded that, compared with the centralized planning method, the reliability indicators of the planning method of this embodiment are all improved. This is because the planning method of this embodiment first divides the cluster, which can ensure the reasonable transmission of node power and increase the probability of island operation, so that the node power supply time is longer and less likely to have power outages, thereby ensuring the reliability of the planning scheme.

[0055] Table 3 Computation time of different algorithms

[0056] It can be concluded from Table 3 that the distribution network planning scheme based on cluster division is 23.21s faster than the centralized planning scheme, which shows that the planning scheme based on cluster division is faster and more superior in calculation speed.

[0057] Scheme P represents a centralized planning scheme, and scheme Q represents a planning scheme based on cluster division. According to the data presented in Table 4, it can be found that compared with the centralized planning scheme, the planning scheme based on cluster division has increased the annual investment cost of photovoltaic energy storage and the annual photovoltaic income by 14.1% and 16.3% respectively. At the same time, the scheme also shows significant advantages in terms of network loss, annual operation and maintenance costs of photovoltaic and energy storage, and user power outage loss costs, which are reduced by 35.9%, 17.7% and 16.8% respectively. This result shows that the planning scheme based on cluster division not only optimizes the economy of distribution network planning and reduces planning costs, but also effectively reduces the losses caused by power outages and improves the reliability of distribution network operation.

[0058] Table 4 Planning costs in different scenarios

[0059] Example 2 A distribution network distributed resource robust planning device based on cluster division includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The method in Example 1 is implemented by executing the computer program by the processor.

Claims

1. A robust planning method for distributed resources in distribution networks based on cluster division, characterized by The following steps are involved: S1. Based on cluster division, aiming at the uncertainty of source and load within the cluster, uncertainty adjustment parameters are introduced to characterize the uncertainty range of source and load, and a source-reservoir robust site selection and capacity determination model is established as the upper model, and the objective function and constraint conditions of the upper model are set; S2. Based on cluster planning, introduce reliability indicators, build a reliability indicator calculation model within the cluster as the lower-level model, and set the objective function of the lower-level model; S3, combining the upper model with the lower model to obtain a two-layer joint planning model; S4. Aiming at the island operation scenario under cluster partitioning conditions, a reliability merging optimization method integrating the island operation characteristics is designed, the two-layer joint planning model is optimized, and the analytical expression of the reliability index is obtained; S5. Obtain a joint planning solution for distributed resources in the distribution network by solving the optimized two-layer joint planning model.

2. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 1, characterized in that: In S1, the objective function of the upper model includes the annual investment and operation and maintenance costs of photovoltaics, the annual income of photovoltaics, and the annual investment cost of energy storage. For any divided cluster k, with the goal of addressing and sizing within the cluster, the objective function F of the upper model is established as: ； In the formula, F1 is the annual investment cost of photovoltaics and energy storage in cluster k; F2 is the annual operation and maintenance cost of photovoltaics and energy storage in cluster k; F3 is the annual income of photovoltaics in cluster k.

3. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 2, characterized in that: F1 is represented by: ； Where j is the node index in cluster k, is the set of nodes contained in cluster k, is the number of nodes contained in cluster k; is the investment cost per MW of PV; Plan the PV capacity for node j in cluster k; is the discount rate; is the number of years of photovoltaic discount; The investment cost per MW of energy storage; is the planned energy storage capacity of node j in cluster k; r is the discounted energy storage years; F2 is expressed as: ； In the formula, is the annual fixed maintenance cost per MW PV; is the annual fixed maintenance cost per MW of energy storage; F3 is represented by: ； In the formula, t represents the time, T is the number of time; is the on-grid electricity price of distributed power sources; Subsidized electricity prices for distributed power generation; is the output limit of each megawatt of photovoltaic power under given sunshine conditions at time t; is the active load of node j at time t.

4. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 3 is characterized in that: The constraints of the upper model include: Energy storage constraints: ； In the formula, , are the charge states of the energy storage at node j at time t and time t-1 respectively; , are the minimum and maximum states of charge of energy storage, respectively; It is the energy storage charging and discharging efficiency; is the energy storage charging power of node j at time t; is the energy storage discharge power of node j at time t; For charging 01 variables; is the discharge 01 variable; when It means that the energy storage device is in the charging state at time t. is 0, ensuring that the energy storage device cannot be charged and discharged at the same time; similarly, when When , it means that the energy storage device is in the discharge state at time t. Must be 0; if and , indicating that the energy storage device is not charging or discharging at time t; , They are the maximum power limits for energy storage charging and discharging respectively; Power balance constraints: ； ； In the formula, is the set of upstream nodes of node j, i represents one of the upstream nodes; is the active power flowing in the line between node i and node j at time t; is the resistance of the line between node i and node j; is the square value of the current in the line between node i and node j; is the actual active power generation of the photovoltaic power plant at node j at time t; is the total active power transferred from node j to its downstream nodes at time t; is the set of downstream nodes of node j, and c represents one of the downstream nodes; is the reactive power flowing in the line between node i and node j at time t; is the reactance of the line between nodes i and j; is the reactive load of node j at time t; is the actual reactive power generation of the PV of node j at time t; is the total reactive power transferred from node j to its downstream nodes at time t; The photovoltaic capacity constraints allowed to be installed on each node in the cluster are: ； In the formula, is the photovoltaic capacity of node j; is the maximum value of the load demand of node j in the cluster; Power flow constraints: ； ； ； In the formula, is the square of the voltage value of node i at time t; , They are the upper and lower limits of the voltage square respectively; is the maximum current limit; Uncertain constraints: ； In the formula, U represents the box-type uncertain set, and u is the specific variable vector in the uncertain set, that is, ; , , At this time, they represent the uncertain variables of active load power, reactive load power, and photovoltaic output of node j in cluster k respectively; , , are the predicted values ​​of active load power, reactive load power, and photovoltaic output of node j in cluster k, respectively; , , They are active load fluctuation amplitude, reactive load fluctuation amplitude, and photovoltaic output fluctuation amplitude respectively; , , are the uncertain adjustment parameters introduced for active load power, reactive load power and photovoltaic output, respectively, and their values ​​are all between [-1, 1], indicating the proportion of the active load power, reactive load power and photovoltaic output fluctuating upward and downward within the fluctuation range during the simulation period; , , are the fluctuation deviations of active load power, reactive load power and photovoltaic output of node j in cluster k respectively.

5. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 4 is characterized in that: In S2, the objective function of the lower model is: ； Where F4 is the power outage loss cost of users in cluster k; is the average load value of node j; is the failure rate of the jth node; is the average power outage time of the jth node failure; is the loss cost caused by power outage at node j based on the average load of node j; and is a reliability indicator; in: ； ； In the formula, is the total number of nodes connecting branches in cluster k; The probability of running for cluster islands; is the maximum interruption amount of the interruptible average node load for node j; is the maximum interruption time of the interruptible average node load for node j.

6. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 5, characterized in that: In the above-mentioned S4, the reliability merging optimization method of the integrated island operation characteristics is to calculate the reliability index of each component in each branch network to the branch node through the reverse flow of the reliability index; simplify the branch network by the reverse flow merging method, and the network becomes a simplified network after the branch network is equivalent; the simplified network is used to quickly evaluate the impact of the component failure in the main network on the branch node, and the analytical expression of the reliability index of the branch node is obtained.

7. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 6, characterized in that: For case 1 where the cluster tie switch is a circuit breaker and case 2 where the cluster tie switch is an isolating switch, analytical expressions of the reliability index are designed respectively.

8. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 7, characterized in that: The analytical expression of the reliability index is designed as follows: Case 1: The cluster tie switch is a circuit breaker, then the failure rate of the jth node based on the downstream merge is , Average power outage time as follows: ； ； In the formula, is the power outage rate obtained by merging the branch node q as the main network node; is the power outage rate obtained by the reverse flow transmission of branch node q as the main network node; is the power outage rate transmitted from node c to branch node q in reverse flow; is the average power outage time obtained by merging the branch node q as the main network node downstream; is the average power outage time obtained by the reverse flow transmission of branch node q as the main network node; is the average power outage time transmitted from node c to branch node q in reverse flow; Case 2: The cluster tie switch is an isolation switch, then the failure rate of the jth node is , Average power outage time as follows: ； ； Finally, the mathematical expression related to the failure rate of the load point, the duration of each failure and the probability variable of cluster island operation is obtained, which is: and The analytical expression of is: ； ； In the formula, is the failure rate of the jth node obtained by the reverse flow transfer method; is the average power outage time of the jth node obtained by the reverse flow transfer method.

9. The method for robust planning of distributed resources in distribution networks based on cluster division according to claim 8, characterized in that: When a failure occurs in the cluster, if the conditions for island operation cannot be met, the maximum interruption time and maximum interruption load of the interruptible load need to be considered, and the following constraints need to be set: Maximum interruption time constraint: ； In the formula, is the upper limit of the maximum interruption time of the interruptible average node load for node j; Maximum interruption constraint: ； In the formula, is the upper limit of the maximum interruption amount of the interruptible average node load for node j; In addition, power balance constraints, power flow constraints, and uncertainty constraints are set in the same way as the relevant constraints in the upper model.

10. A distributed resource robust planning device for a distribution network based on cluster division, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and the method according to any one of claims 1 to 9 is implemented by executing the computer program by the processor.

Images