Disclosure of Invention
In order to overcome the problems, the invention provides a method for dividing a power distribution network dynamic island containing a distributed power supply, which fully considers the balance problem of DG power generation and load power consumption and dynamically divides the island so as to ensure the safety and reliability of island operation; and the island is fused, the utilization rate of the DG is improved, and the power supply recovery range is further expanded.
The technical scheme of the invention is as follows:
a method for dynamic island division of a power distribution network comprising a distributed power supply comprises the following steps:
acquiring initial data, wherein the initial data comprises fault position information, planned power failure time information, power distribution network parameters during fault, load grade parameters, load power prediction and energy storage charge state in the period;
calculating the feasible solution of the island in the period, taking load nodes connected with the distributed power supply and the energy storage as root nodes according to the initial data, and determining the range of each power circle by adopting a breadth-first search algorithm on the premise of meeting related constraints to obtain the feasible solution of the island in the period;
dividing an island range, and searching a preliminary island division range of the time period corresponding to the maximum equivalent load recovery scheme by adopting a depth-first search algorithm on the premise of meeting related constraints according to the feasible island solution of the time period;
optimizing an island division range, and adjusting and optimizing the initial island division range in the period by adopting an island fusion inspection strategy according to the initial island division range in the period to determine the optimal island division range in the period;
and judging whether the optimal islanding ranges of all the time periods are determined, if not, calculating the energy storage charge state of the next time period, determining the optimal islanding ranges of the next time period until the optimal islanding ranges of all the time periods are determined, bringing the optimal islanding ranges of all the time periods into a dynamic islanding scheme set, outputting the dynamic islanding scheme set, and finishing the islanding.
The specific operation of determining the range of each power circle by using the breadth-first search algorithm is as follows:
setting an array a for indicating whether the node is searched; searching all root nodes at the current moment by taking the load nodes connected with the distributed power supply and the energy storage as root nodes, and recording the root nodes as searched nodes in the array a; taking a distributed power supply and energy storage output connected with a root node as a power circle radius of the root node, and if the power circle radius is larger than the load power of the root node, taking the root node as a feasible node and recording the feasible node into a queue A;
searching all lower-layer nodes adjacent to the root node in the network, and recording the lower-layer nodes as searched nodes in the array a; if the sum of the load power of a certain lower-layer node and the load power of a node connected to the upper layer of the node is smaller than the power circle radius of the root node to which the node belongs, recording the node as a feasible node into the queue A, otherwise, not operating the node and stopping searching the node connected to the lower layer of the node;
starting from the lower-layer node recorded in the queue A, searching all nodes of the lower layer which are not searched, recording the nodes of the lower layer as searched nodes in the array a, recording the nodes into the queue A if the sum of the load power of a certain node and the nodes connected with the upper layer is smaller than the power circle radius of the root node to which the node belongs, and otherwise, not operating the nodes and stopping searching the nodes connected with the lower layer; and by analogy, searching layer by layer until all nodes in the network are searched, wherein the nodes contained in the queue A are feasible solutions of island division.
The depth-first search algorithm determines a preliminary island division range by taking the maximum equivalent load recovery quantity as a target; the objective function established with the maximum equivalent load recovery amount is as follows:
in the formula, Nt,LRepresenting the total number of load nodes contained in the island system at the time t; x is the number ofiTaking a value of 1 or 0 when xiWhen the value is equal to 1, indicating that the node i participates in the islanding, otherwise, not participating in the islanding; c. CiReflecting the load weight of the node i, wherein the higher the priority level of the load is, the larger the weight is; pL,i,tThe active power loaded by the node i at the moment t.
When the preliminary island division range of the time period corresponding to the maximum equivalent load recovery scheme is searched by adopting the depth-first search algorithm, various electrical constraint conditions are considered, wherein the constraint conditions of the island division comprise,
firstly, node voltage and branch current constraint:
in the formula: v
i,tIs the voltage at node i at time t,V
i max、V
i minrespectively representing the upper limit and the lower limit of the voltage of a node i; I.C. A
ij,tFor the current flowing on branch i-j during time t,
allowing the maximum current for branch i-j;
node power balance constraint:
in the formula: vi,t、Vj,tThe voltages of the nodes i and j in the period t; alpha is alphaij,tFor the switch state of the line i-j in the period t, taking 0 to represent that the line i-j switch is disconnected, and taking 1 to represent that the line i-j switch is closed; p isis,t、Qis,tRespectively injecting active power and reactive power into a node i at the moment t; gij、BijFor the conductance and susceptance, delta, of the branches i-jij,tThe voltage phase angle difference of the branch i-j is t time period; c (i) is a node set connected with the node i;
③ DG Power constraint:
in the formula: p
DG,i,t、Q
DG,i,tActive output and reactive output of DG at a node i in a period t;
the upper limit of active output and reactive output of DG at a node i in the t period;
the lower limit of active output and reactive output of DG at a node i in the t period;
network structure constraint:
in the formula: f. ofdiFor the virtual load of node i, the units 1, f can be takenij,tIs the virtual flow passing through the branch i-j in the period t, NbIs the number of branches, NnIs the number of nodes, NsThe number of power supplies;
energy storage charging and discharging state and power constraint:
in the formula:
respectively representing 0-1 variables of the charge and discharge state of the energy stored at the node i in the period t;
respectively representing the maximum power of charging and discharging of the energy stored at the node i;
representing the charging and discharging power of the energy storage at the node i in the t period;
sixthly, energy storage residual capacity constraint:
in the formula:
the residual capacity of energy stored at the node i in the period t;
and
maximum and minimum capacity limits of energy storage at node i; eta
ch、η
disRespectively the charge and discharge efficiency of the stored energy;
seventh, capacitor switching constraint:
in the formula:
for the reactive compensation capacity of the capacitor at node i during the period t,
representing the compensation capacity put into a single capacitor,
the number of capacitors switched at node i for the t period,
the total number of capacitors available for switching at node i.
The island fusion inspection strategy is adopted to adjust and optimize the island fusion inspection strategy, and the specific operation of determining the optimal island division range in the period is as follows:
judging whether the primary island division range has an intersection or not, and clearing the intersection island division result to obtain an equivalent node; the removing specifically includes that when island division results corresponding to two adjacent distributed power supplies have an intersection, all nodes on a communication path between distributed power supply access nodes in an island are combined into an equivalent node, distributed power supplies in the two islands are combined into an equivalent distributed power supply, and if a branch exists on the communication path between the distributed power supply access nodes, the branch is directly connected with the equivalent node;
determining a preliminary island division range according to breadth-first search and depth-first search by taking the equivalent node as a root node, judging whether the preliminary island division range has an intersection, and removing island division results with the intersection to obtain the equivalent node; repeating the steps until all the islands have no intersection, and obtaining an optimal island division range without intersection;
fusing any two adjacent islands according to the optimal island division range, and outputting an island division result in the time period; the merging is to merge all nodes in the line between the two island ranges into the island range, if the merged island meets the constraint condition of island division, island merging operation is carried out, otherwise, island merging is cancelled, and the optimal island division range of the time period is obtained.
The invention has the following beneficial effects:
1. the island division method considers the randomness and the volatility of the power grid, dynamically divides the power grid according to the electric quantity balance of the generated energy of the distributed power supply and the electric quantity of the load power consumption, obtains island division ranges in different time periods, and ensures the safety and the reliability of island operation.
2. According to the island division method, the island division result is adjusted and optimized by utilizing a fusion inspection strategy in consideration of the condition of island fusion, islands with intersection in island division ranges are fused, and then adjacent islands are fused, so that the utilization rate of the distributed power supply is improved.
Detailed Description
The invention is described in detail below with reference to the figures and the specific embodiments.
A dynamic islanding method for a power distribution network comprising a distributed power supply comprises the following steps:
acquiring initial data, wherein the initial data comprises fault position information, planned power failure time information, power distribution network parameters during fault, load grade parameters, load power prediction and energy storage charge state in the period; the fault position information is network information of a power grid fault; the planned power failure time information comprises total power failure time of a fault plan, a planned average time period and the current time; the load type represents the priority level of the load, and the load type is divided into a first type load, a second type load and a third type load in the embodiment of the invention;
calculating the feasible solution of the island in the period, taking load nodes connected with the distributed power supply and the energy storage as root nodes according to the initial data, and determining the range of each power circle by adopting a breadth-first search algorithm on the premise of meeting related constraints to obtain the feasible solution of the island in the period;
dividing an island range, and searching a preliminary island division range of the time period corresponding to the maximum equivalent load recovery scheme by adopting a depth-first search algorithm on the premise of meeting related constraints according to the feasible island solution of the time period;
optimizing an island division range, and adjusting and optimizing the initial island division range in the period by adopting an island fusion inspection strategy according to the initial island division range in the period to determine the optimal island division range in the period;
and judging whether the optimal islanding ranges of all the time periods are determined, if not, calculating the energy storage charge state of the next time period, determining the optimal islanding ranges of the next time period until the optimal islanding ranges of all the time periods are determined, bringing the optimal islanding ranges of all the time periods into a dynamic islanding scheme set, outputting the dynamic islanding scheme set, and finishing the islanding.
The specific operation of determining the range of each power circle by using the breadth-first search algorithm is as follows:
setting an array a for indicating whether the node is searched; searching all root nodes at the current moment by taking the load nodes connected with the distributed power supply and the energy storage as root nodes, and recording the root nodes as searched nodes in the array a; taking a distributed power supply and energy storage output connected with a root node as a power circle radius of the root node, and if the power circle radius is larger than the load power of the root node, taking the root node as a feasible node and recording the feasible node into a queue A;
searching all lower-layer nodes adjacent to the root node in the network, and recording the lower-layer nodes as searched nodes in the array a; if the sum of the load power of a certain lower-layer node and the load power of the node connected to the upper layer is smaller than the power circle radius of the root node to which the node belongs, recording the node as a feasible node into the queue A, otherwise, not operating the node and stopping searching the node connected to the lower layer;
starting from the lower node recorded in the queue A, searching all nodes of which the lower layer is not searched, recording the nodes of the lower layer as searched nodes in the array a, recording the nodes of the lower layer into the queue A if the sum of the load power of a certain node and the nodes connected to the upper layer is smaller than the power circle radius of the root node to which the node belongs, and otherwise, not operating the nodes and stopping searching the nodes connected to the lower layer; and by analogy, searching layer by layer until all nodes in the network are searched, wherein the nodes contained in the queue A are feasible solutions of island division.
The depth-first search algorithm determines a preliminary island division range by taking the maximum equivalent load recovery quantity as a target; the objective function established with the maximum equivalent load recovery amount is as follows:
in the formula, Nt,LRepresenting the total number of load nodes contained in the island system at the time t; x is the number ofiTaking a value of 1 or 0 when xiWhen the number is 1, indicating that the node i participates in the islanding, otherwise, not participating in the islanding; c. CiThe load weight of the node i is reflected, the higher the priority level of the load is, the larger the weight is, the first-class load weight is 1, the second-class load weight is 0.5, and the third-class load weight is 0.1 in the embodiment of the invention; pL,i,tThe active power loaded by the node i at the moment t.
When the preliminary island division range of the time period corresponding to the maximum equivalent load recovery scheme is searched by adopting the depth-first search algorithm, various electrical constraint conditions are considered, wherein the constraint conditions of the island division comprise,
firstly, node voltage and branch current are constrained, and after a load is recovered and connected to the grid in an island and reconstruction mode, in order to ensure that a system can normally and stably operate, the node voltage and the branch current in a network need to meet certain constraint conditions, namely:
in the formula: v
i,tIs the voltage of node i at time t, V
i max、V
i minRespectively representing the upper limit and the lower limit of the voltage of a node i; i is
ij,tFor the current flowing on branch i-j during time t,
allowing the maximum current for branch i-j;
secondly, node power balance constraint is carried out, and the key of stable operation of an island and an active power distribution network after reconstruction is to meet the requirement of power balance. As known from kirchhoff's law, the sum of the powers flowing into a node must be equal to the sum of the powers flowing out of the node, and therefore the power balance constraint of the node should be satisfied, namely:
in the formula: vi,t、Vj,tThe voltages of the nodes i and j in the period t; alpha (alpha) ("alpha")ij,tFor the switch state of the line i-j in the period t, taking 0 to represent that the line i-j switch is disconnected, and taking 1 to represent that the line i-j switch is closed; pis,t、Qis,tRespectively injecting active power and reactive power into a node i at the moment t; gij、BijFor the conductance and susceptance, delta, of the branches i-jij,tThe voltage phase angle difference of the branch i-j in the period t; c (i) is a node set connected with the node i;
DG power restraint, because wind-powered electricity generation and photovoltaic's the output has stronger intermittent type nature and volatility, want to make the system can the steady operation after island division and network reconstruction, should satisfy DG power restraint, promptly:
in the formula: p
DG,i,t、Q
DG,i,tActive output and reactive output of DG at a node i in a period t;
the upper limit of active output and reactive output of DG at a node i in the t period;
the lower limit of active output and reactive output of DG at a node i in the t period;
and fourthly, network structure constraint, wherein the active power distribution network is generally in closed-loop design and open-loop operation. Therefore, in the fault recovery process, considering the fault recovery strategy that island division is matched with network reconstruction, the active power distribution network needs to satisfy connectivity constraint and radial constraint, that is:
in the formula: f. ofdiFor the virtual load of node i, the units 1, f can be takenij,tIs the virtual flow passing through the branch i-j in the period t, NbIs the number of branches, NnIs the number of nodes, NsThe number of power supplies;
energy storage charge-discharge state and power constraint, energy storage devices can be used in the process of island division and network reconstruction, but the charge-discharge power of energy storage is not greater than the limit value, so that the energy storage charge-discharge state and power constraint are required to be met, namely:
in the formula:
respectively representing 0-1 variables of the charge and discharge state of the energy stored at the node i in the period t;
respectively representing the maximum power of charging and discharging of the energy stored at the node i;
representing the charging and discharging power of the energy storage at the node i in the t period;
sixthly, energy storage residual capacity constraint, wherein the energy storage has certain capacity limitation, cannot be overcharged or discharged, and should meet the energy storage residual capacity constraint, namely:
in the formula:
the residual capacity of energy stored at the node i in the period t;
and
maximum and minimum capacity limits of energy storage at node i; eta
ch、η
disRespectively the charge and discharge efficiency of the stored energy;
and seventhly, carrying out capacitor switching constraint, wherein in order to meet reactive power requirements of loads during the operation of an island of the active power distribution network and an under-voltage problem caused by reactive power shortage after network reconstruction, a capacitor bank needs to be switched for reactive power compensation in the process of recovering the island and reconstruction faults, and the capacitor switching constraint is met, namely:
in the formula:
for the reactive compensation capacity of the capacitor at node i during the period t,
representing the compensation capacity put into a single capacitor,
the number of capacitors switched at node i for the t period,
the total number of capacitors available for switching at node i.
The island fusion inspection strategy is adopted to adjust and optimize the island fusion inspection strategy, and the specific operation of determining the optimal island division range in the period is as follows:
judging whether the preliminary island division range has intersection or not, and removing the island division result with the intersection to obtain an equivalent node; the removing specifically includes that when island division results corresponding to two adjacent distributed power supplies have an intersection, all nodes on a communication path between distributed power supply access nodes in an island are combined into an equivalent node, distributed power supplies in the two islands are combined into an equivalent distributed power supply, and if a branch exists on the communication path between the distributed power supply access nodes, the branch is directly connected with the equivalent node;
determining a preliminary island division range according to breadth-first search and depth-first search by taking the equivalent node as a root node, judging whether the preliminary island division range has an intersection, and removing island division results with the intersection to obtain the equivalent node; repeating the steps until all the islands have no intersection, and obtaining an optimal island division range without intersection;
fusing any two adjacent islands according to the optimal island division range, and outputting an island division result in the time period; the merging is to merge all nodes in the line between the two island ranges into the island range, if the merged island meets the constraint condition of island division, island merging operation is carried out, otherwise, island merging is cancelled, and the optimal island division range of the time period is obtained.
In an embodiment of the present invention, in an island fusion process when there is an intersection between islands, referring to fig. 3, the island results corresponding to two adjacent distributed power sources have an intersection, nodes (fig. 3, node 6 to node 11) on communication paths between distributed power source access nodes in the island results are merged into an equivalent node (fig. 3, node E), two distributed power sources (fig. 3, DG1, and DG2) are merged into an equivalent distributed power source (fig. 3, DG1+ DG2), and a branch (fig. 3, nodes 15, 16, and 17) exists on a communication path between two distributed power source access nodes, and then the branch is directly connected to the equivalent node.
In an embodiment of the present invention, when there is no intersection in the island division, the island fusion process is, referring to fig. 4, merging all nodes (fig. 4, node 5 to node 11) in the line between any two adjacent island ranges into one island range, if the island meets the constraint condition of the island division, performing island fusion operation, otherwise, canceling the island fusion.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all equivalent structures made by using the contents of the specification and the drawings of the present invention or directly or indirectly applied to other related technical fields are included in the scope of the present invention.