CN110718938B - Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm - Google Patents
Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm Download PDFInfo
- Publication number
- CN110718938B CN110718938B CN201911152017.8A CN201911152017A CN110718938B CN 110718938 B CN110718938 B CN 110718938B CN 201911152017 A CN201911152017 A CN 201911152017A CN 110718938 B CN110718938 B CN 110718938B
- Authority
- CN
- China
- Prior art keywords
- distributed power
- distribution network
- power supply
- node
- load
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
- H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method for reconstructing a distribution network containing a high-proportion distributed power supply based on a Primem algorithm, which comprises the following steps: s1, acquiring output force and load conditions of a distributed power supply in the power distribution network; s2, constructing a demand response model for load and distributed power output adjustment, and adjusting the load and the distributed power output; s3, constructing a reconfiguration model of the distribution network containing the high-proportion distributed power supply; s4, solving the reconstruction model of the distribution network containing the high-proportion distributed power supply by adopting a Primem binary particle swarm algorithm to obtain a reconstruction result. According to the method, a network reconstruction scheme is obtained by constructing a distribution network reconstruction model containing a high-proportion distributed power supply and solving the model by using a binary particle swarm algorithm of Primem. The power distribution network operated according to the scheme is more economic, safe and reliable, and has important practical significance for the sustainable development of the power grid.
Description
Technical Field
The invention relates to a reconstruction method of a distribution network containing a high-proportion distributed power supply based on a Primem algorithm.
Background
With the rapid development of renewable energy power generation, due to considerable environmental benefits, more and more renewable energy-based distributed power sources are connected into a power distribution network, and after a large number of distributed power sources are connected into the power distribution network, the power fluctuation is large, randomness and instability exist, an original power grid topological structure may cause the active power loss and the voltage deviation of the system to be increased, and the economical efficiency of system operation cannot be well met.
Access to a high proportion of distributed power sources is a significant challenge for power systems from a power balance perspective. In the existing network reconstruction method, a reconstruction model is generally selected to have the minimum network loss, the minimum switching operation times, the minimum voltage deviation and the like, and solving algorithms are divided into three types, namely a mathematical optimization algorithm, a heuristic algorithm and an artificial intelligence algorithm. The mathematical optimization algorithm is a mathematical theory, has universality and can obtain an optimal solution theoretically, however, in terms of practical application, the calculation amount of the mathematical optimization algorithm is explosively increased along with the increase of the network order, so that the mathematical optimization algorithm is difficult to be applied to practice. Typical heuristic algorithms are the bypass switching method and the optimal flow pattern method. Compared with a mathematical optimization algorithm, the heuristic algorithm has the advantages that the calculation amount is greatly improved, certain calculation is still needed, the requirement of a branch exchange method on an initial topological structure is high, the optimal flow mode method needs to perform load flow calculation repeatedly, and the calculation amount is also large. Compared with the two types of reconstruction algorithms, the gradually developed artificial intelligence algorithm has the advantages of strong optimizing capability, fast calculation and the like, and is the most widely applied reconstruction algorithm at present. However, in the existing research, a network reconstruction model and an algorithm for a common power distribution network are generally used, the algorithm optimizing efficiency is poor, the running time is long, and the influence of a distributed power supply is not fully considered. Therefore, the research on the network reconstruction method considering the access of the high-proportion distributed power supply to the distribution network has important significance for the safe and stable operation of the power system.
Disclosure of Invention
The invention aims to solve the problems and provides a method for reconstructing a distribution network containing a high-proportion distributed power supply based on a Primem algorithm, which considers the situation that the distributed power supply accesses the distribution network.
In order to achieve the purpose, the technical scheme of the invention is as follows:
the method for reconstructing the distribution network containing the high-proportion distributed power supply based on the Primem algorithm comprises the following steps:
s1, acquiring output force and load conditions of a distributed power supply in the power distribution network;
s2, constructing a demand response model for load and distributed power output adjustment, and adjusting the load and the distributed power output;
s3, constructing a reconfiguration model of the distribution network containing the high-proportion distributed power supply;
s4, solving the reconstruction model of the distribution network containing the high-proportion distributed power supply by adopting a Primem binary particle swarm algorithm to obtain a reconstruction result.
Further, the distributed power output in step S1 includes a fan power output and a photovoltaic power output, the fan power output is calculated by a fan output model, and a calculation formula of the fan output model is as follows:
wherein, PWTOutputting active power of the fan; pr-WTOutputting rated active power of the fan; v. ofciThe cut-in wind speed of the fan; v. ofrIs the rated wind speed of the fan; v. ofcoThe cut-out wind speed of the fan;
the photovoltaic power generation output is calculated by a photovoltaic output model, and the calculation formula of the photovoltaic output model is as follows:
wherein, PPVIs the active power output of the photovoltaic; n ispvIs the number of photovoltaic cells; pr-PVIs the rated active power output of a photovoltaic unit; rrIs the nominal light intensity; k is the power temperature coefficient; t iscAnd TrActual temperature and standard temperature.
Further, the step S2 of constructing the demand response model of load and distributed power output adjustment includes the following steps:
s21, constructing a load demand response model based on price, wherein the formula is as follows:
PL＝αc+β,PL∈[PLmin,PLmax]；
wherein, PLIs the value of the load imposed, alpha and beta are the price elastic coefficients, PLminAnd PLmaxIs the upper and lower limits of the load demand response;
s22, constructing a demand response model for the output adjustment of the distributed power supply, wherein the formula is as follows:
wherein P isDGIn order to obtain the output value of the distributed power supply,
primitive generation for obeying distribution of distributed power supply at current momentForce value, PDGminAnd PDGmaxThe upper limit and the lower limit of the distributed power supply participating in the demand response are respectively, lambda is a demand elasticity coefficient, and lambda is more than 0;
s23, combining the steps S21 and S22 to construct a demand response model of load and distributed power output regulation, wherein the formula is as follows:
wherein the content of the first and second substances,
is the original load value at the present moment without considering the demand characteristics, and γ and μ are the demand elasticity coefficients.
Further, the reconstruction model of the distribution network including the high-proportion distributed power supply in the step S3 adopts a pareto multi-objective optimization method, which includes three objective functions and three constraint conditions at the same time; the objective function is total loss of the network, voltage deviation amount and system reliability; the constraint conditions comprise the maximum capacity of the line allowed to flow, the node voltage value range and the distribution network topological structure.
Further, the calculation formula of the objective function as the total loss of the network is as follows:
wherein, PijAnd QijRespectively, the active power flow and the reactive power flow, U, of the node i side of the line ijiIs the voltage of node i, RijIs the resistance of line ij;
the calculation formula of the objective function as the voltage deviation value is as follows:
wherein, UiIs the actual voltage of node i, UNThe voltage is a rated voltage, n is the number of system nodes, and delta U represents a voltage deviation value level of the whole, and the smaller the value is, the closer the network voltage is to the rated value is;
the calculation formula of the objective function as the system reliability is as follows:
wherein ASAI is the average power supply availability, NiIs the number of users of the load point i, UiIs the annual average outage time for load point i, and R is the set of all load points of the system.
Further, the constraint condition is that the maximum capacity allowed to flow through the line is calculated by the following formula:
wherein S isijmaxIs the maximum capacity, P, that line ij is allowed to flow throughijAnd QijIs the actual power flow of line ij;
the constraint condition is that a calculation formula of the node voltage value range is as follows:
0.95UN≤Ui≤1.05UN；
wherein, UiIs the actual voltage value of node i, UNIs a rated voltage value;
the constraint condition is a power distribution network topological structure, and the power distribution network topological structure is required to meet the requirement of a radiation network structure.
Further, a topological structure of the power distribution network is generated by adopting a Primem algorithm to enable the topological structure to meet a radiation network structure; which comprises the following steps:
s31, preparing a node vector and a branch vector, and respectively storing the node vector, the node obtained by the branch vector and the branch;
s32, storing the network head node into the node vector;
s33, traversing all branches, judging whether one of the nodes at the first end and the last end of the branch is in the node vector and the other node is not in the node vector, if so, storing the branch; if not, continuing to search for the next branch until all branches are traversed;
s34, selecting a branch according to the weight of the branch in the saved branches;
s35, storing the branch selected in the step S34 into branch vectors, and storing nodes which are not in the node vectors into the node vectors;
s36, judging whether the number of nodes in the node vector is equal to the total number of nodes in the network, if so, ending, wherein all branches stored in the branch vector are all branches of the obtained radiation network; if not, the process returns to step S33.
Further, the primm binary particle swarm algorithm in the step S4 includes two particles, one is a global optimal particle, the other is an individual optimal particle, each particle has two attributes of a speed and a position, the speed represents the direction and the size of the next movement, and the position represents the optimal solution to be solved; the velocity and position of the particle are updated by the formula:
wherein the content of the first and second substances,
is the velocity vector of the ith particle for the kth iteration;
is the position vector of the kth iteration of the ith particle; pi kIs the optimal position of the ith particle in all its moved positions after the kth iteration;
the optimal position of all the moved positions of all the particles is the k-th iteration; ρ is the coefficient of inertia which determines how much the last velocity of the particle affects the next velocity; c. C1And c2Acceleration coefficients, which respectively influence the influence of the individual optimal position and the global optimal position on the particle velocity; r is1And r2Is at [0, 1 ]]Random real numbers of inner distribution;
is the position information of the d-dimension of the ith particle in the kth iteration,
is the speed of the d-dimension of the ith particle in the k-1 iteration, r is [0, 1 ]]A random number in between.
Further, in order to embody the original randomness in the binary particle swarm algorithm, the result of multiplying the function value S (x) by the random number is used as the weight of the branch; the formula is as follows:
wherein the content of the first and second substances,
is the branch weight, r, of the kth iteration of the ith branch of the ith particlewIs a [0, 1 ]]A random number in between.
Further, in order to prevent the binary particle swarm algorithm from falling into the local optimum in the optimization process, in the iteration, after the recorded global optimum value is maintained for a certain number of times, the binary particle swarm algorithm is judged to fall into the local optimum, at the moment, the speed of all particles is randomly initialized, and the particles are randomly distributed after the next iteration; the velocity random initialization formula of the particles is:
wherein the content of the first and second substances,
is the speed, V, of the kth iteration of the d-dimension of the ith particlemaxIs the upper speed limit in the S (x) function, rand is [0, 1 ]]A random number in between.
Compared with the prior art, the invention has the advantages and positive effects that:
according to the method, a demand response model for adjusting the output of the load and the distributed power supply is established, the problem that the load and the distributed power supply are not matched can be well solved, and the method has a key effect on the operation of a power distribution network containing a high-proportion distributed power supply; according to the method, a network reconstruction scheme is obtained by solving a model by using a Primem binary particle swarm algorithm with the goals of minimum total network loss, minimum total voltage deviation and maximum average power supply availability, power balance and safe voltage and current stability as constraint conditions and a high-proportion distributed power supply distribution network reconstruction model; the power distribution network operated according to the scheme is more economic, safe and reliable, and has important practical significance for the sustainable development of the power grid.
On the other hand, the Primem binary particle swarm algorithm can effectively avoid the generation of an infeasible solution, so that the model solving efficiency and the solving time are greatly improved, compared with the prior art, the model optimizing method has the advantages of high optimizing efficiency, high calculating speed and strong optimizing capability, and can be well adapted to the random instability of a distributed power supply for a power distribution network containing a high-proportion distributed power supply; and by adopting a timely disturbance strategy, the generation of a local optimal solution can be avoided, so that the final result tends to a global optimal solution, and the optimization capability is effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a block diagram of the framework of the present invention;
fig. 2 is a network structure diagram of an IEEE33 node system in an embodiment.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived from the embodiments of the present invention by a person skilled in the art without any creative effort, should be included in the protection scope of the present invention.
The invention provides a method for reconstructing a distribution network containing a high-proportion distributed power supply based on a binary particle swarm algorithm of Primem. The technical scheme of the invention is shown in figure 1, and comprises the following steps:
s1: and acquiring the output and load conditions of the distributed power supply.
S2: and constructing a demand response model for load and distributed power output adjustment, and adjusting the load and the distributed power output.
S3: and constructing a reconstruction model of the distribution network containing the high-proportion distributed power supply.
S4: and solving the model by adopting a Primem binary particle swarm algorithm according to the network reconstruction model to obtain a network reconstruction result.
S1: acquiring the output and load conditions of the distributed power supply;
the distributed power supply mainly considers fan and photovoltaic power generation. The output of the fan and the photovoltaic power generation is mainly influenced by environmental factors such as wind speed, illumination intensity and the like. The value of the environmental variable can be obtained according to the prediction of a meteorological department.
The fan output model adopts a Weibull model, and the model is shown as the following formula:
in the formula PWTOutputting active power of the fan; pr-WTOutputting rated active power of the fan; v. ofciThe cut-in wind speed of the fan; v. ofrIs the rated wind speed of the fan; v. ofcoIs the cut-out wind speed of the fan.
The photovoltaic output model adopts Beta distribution, which is shown as the following formula:
in the formula, PPVIs the active power output of the photovoltaic; n ispvIs the number of photovoltaic cells; pr-PVIs the rated active power output of a photovoltaic unit; rrIs the nominal light intensity; k is the power temperature coefficient; t iscAnd TrActual temperature and standard temperature.
The load situation is obtained from a typical daily load curve.
S2, constructing a demand response model for load and distributed power output adjustment, and adjusting the load and the distributed power output;
with the access of a large number of distributed power supplies to a power distribution network, the load flow distribution of the system can be changed greatly, and therefore more and more loads adopt a demand response model of time-of-use electricity price regulation to improve the running state of the system. The price-based load demand response model is shown as follows:
PL＝αc+β,PL∈[PLmin,PLmax]； (3)
wherein P isLIs the value of the load exerted, alpha andbeta is the price elastic coefficient, PLminAnd PLmaxAre the upper and lower limits of the load demand response. Generally, α < 0 and β > 0 are always satisfied according to the demand response characteristics of the load.
The output of a distributed power supply depends on the natural environment and is therefore often used in conjunction with an energy storage device in order to turn it into a controllable element. The regulation function of the stored energy can control the overall output of the distributed power supply and the energy storage device. In order to improve the running state of the system, when the load is small, if the output of the distributed power supply is large, the output is correspondingly reduced; when the load is large, if the output force of the distributed power supply is small, the output force should be increased correspondingly. Therefore, the demand response model for distributed power output regulation is established as follows:
wherein P isDGIn order to obtain the output value of the distributed power supply,
an original force output value, P, for the distributed power supply obeying the distribution at the current momentDGminAnd PDGmaxThe upper limit and the lower limit of the distributed power supply participating in the demand response are respectively, lambda is a demand elasticity coefficient, and lambda is larger than 0.
Considering that the load and the distributed power output are in an accompanying relationship, when the load is large, the distributed power output is correspondingly increased to meet the local load requirement, the output of the load also needs to consider the condition of the distributed power output, and the requirement response models of the load and the distributed power output are related to the difference value between the two, so the load and the distributed power output model considering the requirement response can be expressed by the following sub-expressions:
wherein
Is the original load value at the present moment without considering the demand characteristics, and gamma and mu are demand elasticity coefficients, both positive numbers.
When the distributed power supply output is smaller than the load, the load is correspondingly reduced, and the output is increased through the energy storage device, so that the difference value between the total output of the distributed power supply and the stored energy and the load is reduced, the electric energy absorbed from the outside in a region is reduced, the trend of a circuit is reduced, and the effect of improving the running state of the system is achieved. When the output of the distributed power supply is larger than the load, the load is increased by a part, the difference between the total output of the distributed power supply and the total output of the stored energy is reduced, and meanwhile, the output of the energy storage device is reduced, and the consumption of the output of the distributed power supply is increased. The two ends are mutually adaptive to balance electric quantity, so that the trend on a line is reduced, and the function of improving the running state of a system is achieved.
S3: constructing a mathematical model of network reconstruction, namely a reconstruction model of a distribution network containing a high-proportion distributed power supply;
for a distribution network containing a high-proportion distributed power supply, due to the access of the high-proportion distributed power supply, the load flow distribution of the original network is changed, and the total loss of the network, the voltage deviation and the system reliability are greatly influenced. Therefore, the three indexes are used as objective functions, a pareto multi-objective optimization method is adopted, and a network reconstruction model containing a high-proportion distributed power supply is established.
1. Objective function
(1) Targeting total loss of the network
The active loss of the network is directly related to the cost of power grid operation, and has important significance, and the target function is shown as the formula (7):
wherein P isijAnd QijRespectively of line ijNode i side active power flow and reactive power flow, UiIs the voltage of node i, RijIs the resistance of line ij.
(2) Targeting the amount of voltage deviation
The voltage deviation describes the voltage level of each node, and the voltage of each node is required to be not beyond a limited range for the power system to operate normally, so after the network is reconstructed, the smaller the deviation between the node voltage and the rated voltage is, the better the deviation is, and the expression is shown as formula (8):
wherein U isiIs the actual voltage of node i, UNIs the rated voltage and n is the number of system nodes. Δ U represents a voltage deviation level of the whole, and a smaller value indicates that the network voltage is closer to the rated value.
(3) Targeting system reliability
The system reliability index is an important index for evaluating the power distribution network, and the system reliability is calculated by adopting the average power supply availability. The calculation formula is shown as formula (9):
wherein ASAI is the average power supply availability, NiIs the number of users of the load point i, UiIs the annual average outage time for load point i, and R is the set of all load points of the system.
2. Pareto multi-objective optimization
The pareto optimal solution is that each step of improvement makes one target better while other targets are guaranteed not to be worsened, and when a state that improvement cannot be continuously made is reached, the pareto optimal solution is called. This improvement method is called pareto improvement. The condition for achieving the pareto optimal solution is judged to be that the marginal costs of a plurality of targets are equal, namely, the cost brought by increasing the value of one target function is equal to the cost reduced by correspondingly reducing the value of one target function. From the achievement condition of the pareto optimal solution, it can be concluded that the pareto optimal solution is a curve on which all points are likely to be optimal solutions. Therefore, the multi-objective optimization problem in power distribution network reconstruction can be well solved by using the pareto multi-objective optimization, and the result of each objective function is considered. When the power distribution network is reconstructed, a better solution meeting the pareto improvement requirement is selected as an optimal solution of the current state by comparing each objective function value of each scheme, and then the next optimization iteration is carried out. And obtaining a global optimal solution by iterating for a certain number of times.
3. Constraint conditions
According to actual operation requirements, the tidal current capacity of each line cannot exceed the maximum allowable flowing capacity of the line, and the node voltage cannot be too high or too low. The specific expressions are shown as formulas (10) and (11):
0.95UN≤Ui≤1.05UN；
(11)
wherein SijmaxIs the maximum capacity, P, that line ij is allowed to flow throughijAnd QijIs the actual power flow, U, of the line ijiIs the actual voltage value of node i, UNIs the rated voltage value.
According to the requirements of 'closed-loop design and open-loop operation' of a power distribution network, the topological structure of the power distribution network needs to meet the structure of a radiation network, so that the generated topological structure needs to be constrained and must be the radiation network. The topological structure is generated by adopting the Primem algorithm in the graph theory, and the generated network topological structure automatically meets the requirement of a radiation network, so that the generation of an infeasible solution is fundamentally prevented, and the calculation speed of the algorithm can be accelerated. The specific method comprises the following steps:
1) a node vector and a branch vector are prepared, and the acquired node and branch are stored respectively.
2) And storing the network head node into the node vector.
3) Traversing all branches, judging whether one of the nodes at the first end and the last end of the branch is in the node vector and the other node is not in the node vector, if so, storing the branch; if not, continuing to search for the next branch until all branches are traversed.
4) And selecting one branch according to the weight of each branch in the stored branches (randomly selecting one branch when the weights are the same).
5) The branch is stored in a branch vector, and nodes not in the node vector are stored in the node vector.
6) Judging whether the number of nodes in the node vector is equal to the total number of nodes in the network, if so, ending, wherein all branches stored in the branch vector are all branches of the obtained radiation network; if not, returning to the step 3).
The radiation net is directly generated by adopting the Primem algorithm, so that the generation of an infeasible solution can be avoided, and the calculation amount is greatly reduced. The problem of low convergence speed of the artificial intelligence algorithm can be well solved.
S4: and solving the model by adopting a Primem binary particle swarm algorithm according to the network reconstruction model to obtain a network reconstruction result.
The method adopts a binary particle swarm optimization algorithm of Primem to solve the model. In the aspect of optimization strategy, in order to prevent the optimization algorithm from falling into a local optimal state, the invention adopts a timely disturbance strategy to enable the algorithm to better search global optimization.
1. Binary particle swarm algorithm for Primem
The particle swarm algorithm uses the interaction of attraction between particles, and a particle with "larger size" can attract other particles to move to its own position, so that a new particle is generated in the process of movement, and a "larger particle" may exist in the new particle, so that the particle swarm moves to the new particle.
In the particle swarm optimization, two 'individual larger' particles are provided, one is a global optimal particle, the other is an individual optimal particle, each particle has two attributes of speed and position, the speed represents the direction and the size of the next movement, and the position represents the solution of the optimization algorithm. The binary particle swarm algorithm is a binary array with positions only consisting of 0 and 1. The velocity and position updating formulas are shown in formulas (12) to (14):
wherein the content of the first and second substances,
is the velocity vector of the ith particle for the kth iteration;
is the position vector of the kth iteration of the ith particle; pi kIs the optimal position of the ith particle in all its moved positions after the kth iteration;
the optimal position of all the moved positions of all the particles is the k-th iteration; rho is an inertia coefficient which determines the influence degree of the previous speed of the particle on the next speed, and is generally 0.3-0.9; c. C1And c2Acceleration coefficients respectively influencing the influence of the individual optimal position and the global optimal position on the speed of the particle are respectively, and the acceleration coefficients are generally 2; r is1And r2Is at [0, 1 ]]Random real numbers of inner distribution;
is the ith particle in the kth iterationThe position information of the d-th dimension, similarly,
is the d-dimension velocity of the ith particle in the k-1 iteration, r is [0, 1 ]]A random number in between. The S (x) function takes into account the speed out of limit and defines the function value as a fixed value when the speed exceeds a maximum value or is less than a minimum value.
Binary particle swarm optimization is more suitable for network reconstruction of a power distribution network because each dimension in the position of a particle corresponds to one line in the network, network reconstruction is reconstruction of line states, and the line states are only two, namely open and closed, and exactly correspond to 0 and 1 in a binary system. Therefore, the invention adopts a binary particle swarm algorithm.
The branches are selected according to the weight of each branch in the primum algorithm, and the size of the weight directly influences the generated network topology. In the binary particle swarm algorithm, an S-function determined by the velocity affects the position information of the particles. Since both of them will affect the generated network topology, the primm binary particle swarm algorithm combines both to complete the control of the network reconstruction.
As can be seen from equations (12) to (14), the larger the particle velocity, the larger the S function value, and theoretically, the position update is biased to take 1; on the other hand, the pram algorithm reselects the line according to the branch weight, the larger the weight is, the more the selection is favored, i.e. the position of the branch is favored to take 1. Therefore, the S function value and the utility of the branch weight value in the Primem algorithm are the same, so that the S function value can be indirectly given to the branch as the weight, and the Primem algorithm and the binary particle swarm algorithm are combined.
In order to embody the original randomness of the binary particle swarm algorithm and enable the searching effect of the algorithm to be higher, the result obtained by multiplying the S function value by the random number is used as the weight of the branch. The specific formula is shown as formula (15):
wherein
Is the branch weight, r, of the kth iteration of the ith branch of the ith particlewIs a [0, 1 ]]A random number in between. According to the formula, the current velocity value of a particle affects the branch weight of the next particle, thereby affecting the position of the particle. In addition, random variables are added, so that the positions of the particles have random characteristics, and better search optimization can be realized.
2. Timely perturbation strategy
In order to prevent the binary particle swarm optimization algorithm from falling into local optimization in the optimization process, some measures are needed to realize global search, and therefore the method adopts a timely disturbance strategy. The timely disturbance is added according to a specific situation as the name implies, specifically, after a recorded global optimum value is maintained for a certain number of times in iteration, the disturbance is judged to possibly fall into local optimum, at the moment, the speed of all particles is randomly initialized, the particles are randomly distributed after the next iteration, and then the individual optimum and the global optimum attract the particles, so that the global search is realized. The velocity random initialization formula is shown as equation (16):
wherein the content of the first and second substances,
is the speed, V, of the kth iteration of the d-dimension of the ith particlemaxIs the upper speed limit in the S function, and rand is [0, 1 ]]A random number in between.
The invention adopts an IEEE33 node system as an example analysis, and the grid structure is shown in figure 2. The total load was 3715kW, 2300 kVar. The voltage reference value was 12.66 kV. And the load flow calculation adopts a forward-backward substitution method. In the Primem binary particle swarm algorithm, the inertia coefficient rho is 0.7, and the acceleration coefficient c1And c2Get 2, VmaxAnd Vmin3.9 and-3.9 are respectively taken, the population number is 30, and the iteration is carried out for 300 times. In the timely disturbance strategy, after the global optimum is maintained for 10 iterations, the randomized disturbance is executed. In the demand response model, the demand elastic coefficients γ and μ take 0.05. When the average power supply availability is calculated, the fault rate of each line is set to be 0.0065, and the average power failure time of each fault is set to be 1 hour. For simple calculation, the distributed power nodes adopt a PQ model, and the power factor is 0.9. The position and installed capacity are shown in table 1 and fig. 2, and the power unit is MW. In fig. 2, the fan is connected to the nodes 7, 8, 13 and 24, and the photovoltaic is connected to the nodes 17, 23 and 29. The fan output and the photovoltaic output change along with time, and the output value of each time period is assumed to be fixed, as shown in table 2; the branch impedance and node loading conditions are shown in table 3.
The distributed power capacity and access point are shown in table 1:
TABLE 1 distributed Power location and Capacity
| Position of | 7 | 8 | 13 | 17 | 23 | 24 | 29 |
| Active power | 0.12 | 0.12 | 0.12 | 0.21 | 0.21 | 0.12 | 0.21 |
The distributed power output over time is shown in table 2:
TABLE 2 distributed power output at different times
| Time of day | Fan blower | Photovoltaic system | Time of day | Fan blower | Photovoltaic system |
| 1 | 0.12 | 0 | 13 | 0.05 | 0.21 |
| 2 | 0.10 | 0 | 14 | 0.06 | 0.21 |
| 3 | 0.09 | 0 | 15 | 0.05 | 0.20 |
| 4 | 0.07 | 0 | 16 | 0.06 | 0.18 |
| 5 | 0.07 | 0 | 17 | 0.08 | 0.13 |
| 6 | 0.07 | 0 | 18 | 0.09 | 0.08 |
| 7 | 0.05 | 0.05 | 19 | 0.08 | 0.05 |
| 8 | 0.08 | 0.08 | 20 | 0.12 | 0 |
| 9 | 0.05 | 0.13 | 21 | 0.11 | 0 |
| 10 | 0.07 | 0.18 | 22 | 0.11 | 0 |
| 11 | 0.08 | 0.19 | 23 | 0.11 | 0 |
| 12 | 0.09 | 0.20 | 24 | 0.09 | 0 |
TABLE 3 Branch impedance and node load conditions
The distribution network is reconstructed according to the points of S2, S3, and S4, taking points of 6, 12, and 20 as examples, and the calculation results are shown in table 4:
table 4 network reconstruction results at different times
As can be seen from table 4, the operation state of the system is kept good after the network reconfiguration by the algorithm of the present invention, regardless of the distributed power source output.
The demand response model for load and distributed power supply output adjustment can well solve the problem that the load and the distributed power supply are not matched, and has a key effect on the operation of a power distribution network with a high-proportion distributed power supply. The Primem binary particle swarm algorithm can effectively avoid the generation of an infeasible solution, thereby greatly improving the model solving efficiency and the solving time, and being well suitable for the randomness instability of a distributed power supply for a power distribution network containing a high-proportion distributed power supply.
According to the method, a network reconstruction scheme is obtained by solving a model by using a binary particle swarm algorithm based on a Primem algorithm and taking the minimum total network loss, the minimum total voltage deviation and the maximum average power supply availability as targets, and taking power balance and voltage and current safety stability as constraint conditions. The power distribution network operated according to the scheme is more economic, safe and reliable, and has important practical significance for the sustainable development of the power grid.
Claims (8)
1. A method for reconstructing a distribution network containing a high-proportion distributed power supply based on a Primem algorithm is characterized by comprising the following steps: the method comprises the following steps:
s1, acquiring output force and load conditions of a distributed power supply in the power distribution network;
in the step S1, the distributed power output includes a fan power output and a photovoltaic power output, the fan power output is calculated by a fan output model, and a calculation formula of the fan output model is as follows:
wherein, PWTOutputting active power of the fan; pr-WTOutputting rated active power of the fan; v. ofciThe cut-in wind speed of the fan; v. ofrIs the rated wind speed of the fan; v. ofcoThe cut-out wind speed of the fan;
the photovoltaic power generation output is calculated by a photovoltaic output model, and the calculation formula of the photovoltaic output model is as follows:
wherein, PPVIs the active power output of the photovoltaic; n ispvIs the number of photovoltaic cells; pr-PVIs the rated active power output of a photovoltaic unit; rrIs the nominal light intensity; k is the power temperature coefficient; t iscAnd TrActual temperature and standard temperature;
s2, constructing a demand response model for load and distributed power output adjustment, and adjusting the load and the distributed power output;
the step S2 of constructing the demand response model of load and distributed power output adjustment includes the following steps:
s21, constructing a load demand response model based on price, wherein the formula is as follows:
PL＝αc+β,PL∈[PLmin,PLmax]；
wherein, PLIs the value of the load imposed, alpha and beta are the price elastic coefficients, PLminAnd PLmaxIs the upper and lower limits of the load demand response;
s22, constructing a demand response model for the output adjustment of the distributed power supply, wherein the formula is as follows:
wherein P isDGIn order to obtain the output value of the distributed power supply,
an original force output value, P, for the distributed power supply obeying the distribution at the current momentDGminAnd PDGmaxThe upper limit and the lower limit of the distributed power supply participating in the demand response are respectively, lambda is a demand elasticity coefficient, and lambda is more than 0;
s23, combining the steps S21 and S22 to construct a demand response model of load and distributed power output regulation, wherein the formula is as follows:
wherein the content of the first and second substances,
is the original load value without considering the demand characteristics at the current moment, and gamma and mu are the demand elasticity coefficients;
s3, constructing a reconfiguration model of the distribution network containing the high-proportion distributed power supply;
s4, solving the reconstruction model of the distribution network containing the high-proportion distributed power supply by adopting a Primem binary particle swarm algorithm to obtain a reconstruction result.
2. The method for reconstructing a distribution network with a high proportion of distributed power sources based on the pram algorithm according to claim 1, wherein: the reconstruction model of the distribution network containing the high-proportion distributed power supply in the step S3 adopts a pareto multi-objective optimization method, and the pareto multi-objective optimization method simultaneously comprises three objective functions and three constraint conditions; the objective function is total loss of the network, voltage deviation amount and system reliability; the constraint conditions comprise the maximum capacity of the line allowed to flow, the node voltage value range and the distribution network topological structure.
3. The method for reconstructing a distribution network with a high proportion of distributed power sources based on the pram algorithm according to claim 2, wherein: the calculation formula of the objective function as the total loss of the network is as follows:
wherein, PijAnd QijRespectively, the active power flow and the reactive power flow, U, of the node i side of the line ijiIs the voltage of node i, RijIs the resistance of line ij;
the calculation formula of the objective function as the voltage deviation value is as follows:
wherein, UiIs the actual voltage of node i, UNThe voltage is a rated voltage, n is the number of system nodes, and delta U represents a voltage deviation value level of the whole, and the smaller the value is, the closer the network voltage is to the rated value is;
the calculation formula of the objective function as the system reliability is as follows:
wherein ASAI is the average power supply availability, NiIs the number of users of the load point i, UiIs the annual average outage time for load point i, and R is the set of all load points of the system.
4. The method for reconstructing the distribution network with the high-proportion distributed power sources based on the pram algorithm as claimed in claim 3, wherein the method comprises the following steps: the calculation formula of the constraint condition as the maximum capacity allowed to flow by the line is as follows:
wherein S isijmaxIs the maximum capacity, P, that line ij is allowed to flow throughijAnd QijIs the actual power flow of line ij;
the constraint condition is that a calculation formula of the node voltage value range is as follows:
0.95UN≤Ui≤1.05UN；
wherein, UiIs the actual voltage value of node i, UNIs a rated voltage value;
the constraint condition is a power distribution network topological structure, and the power distribution network topological structure is required to meet the requirement of a radiation network structure.
5. The method for reconstructing the distribution network with the high proportion of distributed power sources based on the pram algorithm as claimed in claim 4, wherein: generating a power distribution network topological structure by adopting a Primem algorithm to enable the power distribution network topological structure to meet a radiation network structure; which comprises the following steps:
s31, preparing a node vector and a branch vector, and respectively storing the node vector, the node obtained by the branch vector and the branch;
s32, storing the network head node into the node vector;
s33, traversing all branches, judging whether one of the nodes at the first end and the last end of the branch is in the node vector and the other node is not in the node vector, if so, storing the branch; if not, continuing to search for the next branch until all branches are traversed;
s34, selecting a branch according to the weight of the branch in the saved branches;
s35, storing the branch selected in the step S34 into branch vectors, and storing nodes which are not in the node vectors into the node vectors;
s36, judging whether the number of nodes in the node vector is equal to the total number of nodes in the network, if so, ending, wherein all branches stored in the branch vector are all branches of the obtained radiation network; if not, the process returns to step S33.
6. The method for reconstructing a distribution network with a high proportion of distributed power sources based on the pram algorithm according to claim 5, wherein: the primm binary particle swarm algorithm in the step S4 includes two particles, one is a global optimal particle, the other is an individual optimal particle, each particle has two attributes of speed and position, the speed represents the direction and size of the next movement, and the position represents the optimal solution to be solved; the velocity and position of the particle are updated by the formula:
wherein the content of the first and second substances,
is the kth iteration of the ith particleThe velocity vector of (2);
is the position vector of the kth iteration of the ith particle; pi kIs the optimal position of the ith particle in all its moved positions after the kth iteration;
the optimal position of all the moved positions of all the particles is the k-th iteration; ρ is the coefficient of inertia which determines how much the last velocity of the particle affects the next velocity; c. C1And c2Acceleration coefficients, which respectively influence the influence of the individual optimal position and the global optimal position on the particle velocity; r is1And r2Is at [0, 1 ]]Random real numbers of inner distribution;
is the position information of the d-dimension of the ith particle in the kth iteration,
is the speed of the d-dimension of the ith particle in the k-1 iteration, r is [0, 1 ]]A random number in between; vmaxIs the upper speed limit in the S (x) function, and Vmin is the lower speed limit in the S (x) function.
7. The method for reconstructing a distribution network with a high proportion of distributed power sources based on the pram algorithm according to claim 6, wherein: in order to embody the original randomness in the binary particle swarm algorithm, the result of multiplying the function value S (x) by the random number is used as the weight of the branch; the formula is as follows:
wherein the content of the first and second substances,
is the branch weight, r, of the kth iteration of the ith branch of the ith particlewIs a [0, 1 ]]A random number in between.
8. The method for reconstructing a distribution network with a high proportion of distributed power sources based on the pram algorithm according to claim 7, wherein: in order to prevent the binary particle swarm algorithm from falling into the local optimum in the optimization process, in the iteration, after the recorded global optimum value is kept unchanged for a certain number of times, the binary particle swarm algorithm is judged to fall into the local optimum, at the moment, the speed of all particles is randomly initialized, and the particles are randomly distributed after the next iteration; the velocity random initialization formula of the particles is:
wherein the content of the first and second substances,
is the speed of the kth iteration of the d-dimension of the ith particle, and rand is [0, 1 ]]A random number in between.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911152017.8A CN110718938B (en) | 2019-11-22 | 2019-11-22 | Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911152017.8A CN110718938B (en) | 2019-11-22 | 2019-11-22 | Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110718938A CN110718938A (en) | 2020-01-21 |
| CN110718938B true CN110718938B (en) | 2020-12-22 |
Family
ID=69216326
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911152017.8A Active CN110718938B (en) | 2019-11-22 | 2019-11-22 | Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110718938B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114896871B (en) * | 2022-04-22 | 2023-12-22 | 国网江苏省电力有限公司电力科学研究院 | Photovoltaic output decomposition method based on minimum mutual information |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103400207B (en) * | 2013-08-01 | 2014-09-17 | 天津大学 | Operation optimization method for power distribution network comprising schedulable distributed power supply |
| CN110391673B (en) * | 2019-07-03 | 2020-12-15 | 浙江大学 | Multi-period active distribution network reconstruction method considering demand side response under high-proportion wind power |
-
2019
- 2019-11-22 CN CN201911152017.8A patent/CN110718938B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110718938A (en) | 2020-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tolabi et al. | Simultaneous reconfiguration, optimal placement of DSTATCOM, and photovoltaic array in a distribution system based on fuzzy-ACO approach | |
| Hong et al. | Optimal VAR control considering wind farms using probabilistic load-flow and gray-based genetic algorithms | |
| CN111342461B (en) | Power distribution network optimal scheduling method and system considering dynamic reconfiguration of network frame | |
| CN105071433B (en) | A kind of configuration scheme of distributed generation resource | |
| CN114362196B (en) | Multi-time-scale active power distribution network voltage control method | |
| Tang et al. | Study on day-ahead optimal economic operation of active distribution networks based on Kriging model assisted particle swarm optimization with constraint handling techniques | |
| CN109904877B (en) | Distributed wind power plant optimization operation method based on variable power factor | |
| CN106712075A (en) | Peaking strategy optimization method considering safety constraints of wind power integration system | |
| CN111490542B (en) | Site selection and volume fixing method of multi-end flexible multi-state switch | |
| CN110929931A (en) | Power distribution network coordination planning method considering distributed power supply and load time sequence characteristics | |
| Askarzadeh | Solving electrical power system problems by harmony search: a review | |
| Nguyen et al. | Performance Comparison between PSO and GA in Improving Dynamic Voltage Stability in ANFIS Controllers for STATCOM. | |
| Chuang et al. | Design of intelligent control for stabilization of microgrid system | |
| CN110718938B (en) | Method for reconstructing distribution network containing high-proportion distributed power supply based on Primem algorithm | |
| Khorram-Nia et al. | A novel stochastic framework for the optimal placement and sizing of distribution static compensator | |
| CN111193288B (en) | Recovery reconstruction method for distribution network containing distributed power supply based on comprehensive recovery strategy | |
| Zad et al. | Optimal reactive power control of DGs for voltage regulation of MV distribution systems considering thermal limit of the system branches | |
| CN109638837B (en) | New energy grid-connected-oriented dynamic load management method and system | |
| Ranganathan et al. | Self‐Adaptive Firefly‐Algorithm‐Based Unified Power Flow Controller Placement with Single Objectives | |
| Tyagi et al. | Comparative study of IHS and BF-DE algorithm for optimal DG placement | |
| Liu et al. | Optimal power factor regulation of dispersed wind farms under diverse load and stochastic wind conditions based on improved firefly algorithm | |
| CN110930263B (en) | Medium-voltage distribution network short-circuit current calculation method containing photovoltaic power supply and induction motor based on black hole particle swarm algorithm | |
| Ahsan et al. | Performance analysis of different search algorithms in optimization of power system operation | |
| Doagou-Mojarrad et al. | Probabilistic interactive fuzzy satisfying generation and transmission expansion planning using fuzzy adaptive chaotic binary PSO algorithm | |
| Mahider et al. | Optimization of STATCOM PI Controller Parameters Using the Hybrid GA-PSO Algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |