CN110334391B - Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant - Google Patents
Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant Download PDFInfo
- Publication number
- CN110334391B CN110334391B CN201910433214.0A CN201910433214A CN110334391B CN 110334391 B CN110334391 B CN 110334391B CN 201910433214 A CN201910433214 A CN 201910433214A CN 110334391 B CN110334391 B CN 110334391B
- Authority
- CN
- China
- Prior art keywords
- cost
- fan
- clustering
- tree
- algorithm
- 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
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 122
- 238000005457 optimization Methods 0.000 claims abstract description 35
- 238000004364 calculation method Methods 0.000 claims abstract description 27
- 238000010845 search algorithm Methods 0.000 claims abstract description 4
- 238000000638 solvent extraction Methods 0.000 claims abstract description 4
- 238000010586 diagram Methods 0.000 claims description 26
- 230000008569 process Effects 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 11
- 239000003795 chemical substances by application Substances 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 7
- 230000001154 acute effect Effects 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 6
- 238000012067 mathematical method Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 4
- 238000009826 distribution Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 230000002285 radioactive effect Effects 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000002093 peripheral effect Effects 0.000 claims description 3
- 230000006872 improvement Effects 0.000 claims description 2
- 239000011159 matrix material Substances 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000009827 uniform distribution Methods 0.000 claims description 2
- 230000002068 genetic effect Effects 0.000 abstract description 2
- 238000013461 design Methods 0.000 description 7
- 230000003993 interaction Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012938 design process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008685 targeting Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Geometry (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Mathematical Optimization (AREA)
- Game Theory and Decision Science (AREA)
- Computational Mathematics (AREA)
- Computer Hardware Design (AREA)
- Pure & Applied Mathematics (AREA)
- Operations Research (AREA)
- Mathematical Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Structures Of Non-Positive Displacement Pumps (AREA)
Abstract
The invention discloses a multidimensional constraint wind power plant current collection line automatic planning method, which comprises the following steps: 1) Carrying out intelligent partitioning; 2) Optimally planning the cost in the region; 3) The overall cost of the line cross-region planning is optimal; 4) And optimizing the T-connection path. The method solves the problems of large calculation amount and long time of the traditional NP algorithm such as dimension disaster, genetic algorithm, RRT quick search algorithm, ant colony algorithm and the like, carries out regional division and hierarchical individual breaking on the current collecting line planning problems of a large-scale and large-quantity wind power plant, solves the problems of high strength, nonlinearity and high time complexity of the current collecting line planning, ensures that the planning of the wind power plant can be carried out in a considerable and gradual mode, carries out self-optimization under multi-dimensional constraint, and reduces the influence of human factors on the planning result.
Description
Technical Field
The invention relates to the technical field of wind power plant planning, in particular to a multidimensional constraint wind power plant current collection circuit automatic planning method.
Background
At present, wind power plant planning methods at home and abroad are mostly combined with actual needs of specific engineering, designers manually design based on experience, and power collecting line planning needs to be combined with terrain and road planning so as to achieve optimal cost, but the method is limited by the experience of the designers, the planning levels are different, and the power collecting line cost deviation is huge. The current collecting system electrical scheme thus obtained is difficult to obtain good economy. The experience level of designers plays a decisive role in high planning quality, and for large and complex wind power plants, the design period is long, and an optimal collection line topological mode is often difficult to obtain.
The current collection line planning algorithm of each fan manufacturer can only plan a distance optimal topology, but cannot plan a cost optimal topology, a three-dimensional optimal path of a complex terrain is not fully considered, or only radioactive area division is considered but area interaction is ignored, the purpose of cost optimization cannot be achieved frequently in a planning result, and the planning time is basically over half an hour. The algorithm takes the idea of manual design as reference, comprehensively considers various practical problems encountered in the algorithm design process, including rapid sub-area division, three-dimensional path planning of complex terrain, cost-optimized topology, single-double-circuit paths and sub-area interaction, reduces the order for many times, and controls the planning time within 10 minutes.
Wind energy is more and more widely concerned by people as an important renewable energy source, the construction cost of a wind power plant is high, the construction cost of a collecting line is 10% higher in the construction cost of the whole wind power plant, the cost of the collecting system is seriously influenced by the selection of different cable sections of the wind power plant and topological structures of the collecting system, a large-scale wind power plant with more densely arranged fans is often required to be subjected to a large amount of calculation work by designers, the cost of each scheme is compared by comparing different topological structures, and a better scheme is finally selected from the schemes.
Disclosure of Invention
The invention aims to solve the problems of long manual design time and difficult optimization of economy under multi-dimensional constraint, provides an automatic planning method for a multi-dimensional constraint wind power plant current collecting line, solves the problems of large calculation amount and long time of a traditional NP algorithm such as dimension disaster, a genetic algorithm, an RRT (rapid search for transform) algorithm, an ant colony algorithm and the like, plans a large number of large-scale wind power plant current collecting lines, divides regions, breaks through each layer, solves the problems of high strength, nonlinearity and high time complexity of current collecting line planning, enables the planning of the wind power plant to be performed in a considerable and gradual change manner, performs self-optimization under multi-dimensional constraint, and reduces the influence of human factors on a planning result.
In order to achieve the purpose, the technical scheme provided by the invention is as follows: a multi-dimensional constraint wind power plant current collection line automatic planning method comprises the following steps:
1) Intelligent partitioning
The method comprises the following steps of performing regional division, dividing a fan array in a wind power plant into a plurality of radioactive regions by taking a booster station as a center, and limiting the capacity of each region, wherein the planning of a current collecting line of the wind power plant is a multi-dimensional and nonlinear planning problem;
2) Cost-optimized planning in a region
Constructing a graph G (V, E) for each subregion, converting the line planning problem into a graph theory problem, and obtaining the cost optimal topology in each subregion by using a mathematical method of the graph theory;
3) Line cross-regional planning global cost optimization
Increasing feasible paths among the sub-regions for exploration, adding the feasible paths among the regions into a sub-region cost optimal tree, and performing dynamic optimization to obtain a final full-field cost optimal topology;
4) T-connection path optimization
The problem of T-connection cross points is not considered in the obtained cost-optimal topology, and a plurality of T-connection modes exist in practical engineering, so that the obtained line topology needs to be subjected to T-connection optimization to obtain an economical and more optimal collector line connection topology.
In the step 1), intelligent region division is carried out by adopting a fuzzy clustering algorithm, and the specific conditions are as follows:
clustering analysis, which is to classify individuals or objects according to similarity or distance to make the similarity between elements in the same class stronger than that of other classes, aiming at maximizing the homogeneity of elements between classes and the heterogeneity between elements between classes, mainly based on that samples gathered in the same data set should be similar to each other, but samples belonging to different groups should be dissimilar, essentially using computer classification to replace manual classification to perform tedious classification work; the fuzzy clustering algorithm is adopted to divide the area of the whole wind turbine, and in the process of adopting the fuzzy clustering algorithm, the following two limits are required to be carried out in consideration of the actual engineering requirements of the wind power plant:
a. based on the particularity of the wind power plant, radial clustering results with booster stations as the centers are required to be obtained instead of block clustering;
b. due to the limitation of the current-carrying capacity of the cable, the capacity of each current collecting line needs to be limited, so that the capacity of each cluster of fans is detected to determine whether the capacity exceeds the limit;
in the fuzzy clustering algorithm, firstly, the number of clustered clusters, namely the number of sub-regions is determined, and an initial clustering center point plays a decisive role in a final clustering result, and the initial clustering points are selected according to wind power planning practice and are uniformly distributed as far as possible in a dispersing manner so as to conveniently and rapidly obtain reasonable region division, so that the region division is performed after the improvement of the traditional fuzzy clustering algorithm;
wherein the number of sub-regions is determined as follows:
for a wind power plant, the number of sub-regions of a front collecting line is not determined to be unknown in a connection mode, but the number of the divided minimum sub-regions can be obtained through calculation according to the capacity of the wind power plant and the maximum capacity limit of each sub-region, so that the number of the sub-regions is clustered according to the number of the minimum sub-regions, a reasonable iteration stop condition is set, and if clustering cannot be completed, the number of the sub-regions is gradually increased until clustering is successful;
the improved fuzzy clustering algorithm comprises the following steps:
step 1: setting related parameters including the minimum sub-region number min _ group, the maximum sub-region number max _ group, the initial class-core maximum cycle number fuzzy _ max, the maximum cycle number station _ max of the fan not belonging to any class, the cluster-core movement maximum iteration number delt _ max and the tolerance tolerence;
step 2: the method comprises the following steps that exploration is started when the sub-area number group is the minimum set sub-area number min _ group, the cluster centers are grouped at the moment, each sub-area is divided into a cluster, and the current collecting lines planned and formed in each sub-area are called a string or a current collecting line;
and 3, step 3: judging whether the number group of the sub-regions exceeds the maximum value max _ group, if not, executing the following step 4, otherwise, indicating that the clustering fails;
and 4, step 4: dividing the wind field extension into group sub-regions, regarding the double-loop as a sub-region, and since the number of fans in the class, namely the capacity of each sub-region, is not limited in the clustering process, the load flow in the sub-regions needs to be detected after the clustering is finished, and when the capacity limit is exceeded, the clustering needs to be performed again, so that a fuzzy cycle is embedded for the re-circulating clustering under the condition, the fuzzy is initialized to be 0, and the maximum iteration number is fuzzy _ max;
and 5, step 5: judging whether fuzzy is smaller than fuzzy _ max, if yes, executing the following step 6, otherwise, increasing the number of planning sub-regions by group = group +1 and updating, wherein the group is the number of the sub-regions, and executing the step 3;
and 6, a step of: because the distance from the fan to the class center adopts a polar coordinate form, when the included angle between the vector from the booster station to the fan and the vector from the booster station to the class center is more than 90 degrees, the fan does not belong to the class, if the fan does not belong to any class, the initial class center is reselected for clustering, so that a station cycle is embedded for re-clustering under the condition, the station is initialized to be 0, and the maximum iteration number is station _ max;
and 7, step 7: judging whether the station is smaller than station _ max, if so, executing the next step 8, otherwise, circularly accumulating fuzzy = fuzzy +1 to update the cycle times, and executing the step 5;
and 8, step 8: the initial class center of the cluster is obtained by using a roulette algorithm, and in the selection process of the initial class center, group points which are as far away from each other as possible are selected in order to obtain a reasonable clustering effect, so that the iteration steps of the algorithm can be reduced, and the classification result is more uniform; the method mainly comprises the following steps: firstly, randomly selecting a fan point as a first initial cluster center point, then selecting the point farthest from the fan point as a second initial cluster center point, then selecting the point farthest from the first two points as a third initial cluster center point, and so on until group initial cluster center points are selected; after the initial class center is determined, namely all class center points are completely determined, the clustering result is determined, so that the clustering result lacks diversity and even clustering failure can be caused, a roulette algorithm is introduced for the reason, the probability that the points are selected is higher as the distance is longer, instead of selecting the points with the farthest distance by 100%, the uniform distribution of the initial points is ensured, and multiple clustering possible results are provided for clustering;
step 9: calculating the distance d from each fan point to each class center ic I belongs to (1, n _node), c belongs to (1, n _c), i represents the ith fan, n _ node represents the number of fans, c represents the class center, n _ c represents the number of class centers, d ic Represents the distance from the ith fan to the c-th class center, andnormalizing the distances, converting the distances into membership degree matrixes from the fans to various class centers, and dividing fan nodes into different clusters according to the membership degrees; if the fan does not belong to any class, namely the distances from the fan nodes to all the class centers are infinite, the initially selected class centers cannot complete clustering, the initial class centers need to be reselected, the station + + returns to the step 7 above, the initial class centers are reselected for calculation, and otherwise, the step 10 below is continuously executed; the method for calculating the distance from the fan node to the class center comprises the following steps: in order to obtain a radial clustering result, the vertical distance from the fan to the booster station and the connection line of the class center is the distance from the fan to the class center, and the linear distance from the fan to the class center is not used; wherein, the station + + represents that station accumulates one for each cycle for counting;
when the booster station arrives at the class center phasor
Is equal to the amount from the booster station to the fan>
When the included angle is less than or equal to 90 degrees, the distance from the fan to the class center is the distance from the fan to the phase quantity->
Perpendicular distance d of ic = d, if the angle is greater than 90 °, then>
And &>
Is greater than 90, the distance from the fan to the class center is set as d jc ＝∞；
The membership of the fuzzy clusters is calculated as follows:
d ic -is the distance from the ith fan to the c-th centroid;
n _ c-represents the number of centroids;
m-is a weighted index;
member ic -is the membership of the ith fan to the c class center;
for a certain fan, dividing fan nodes into clusters with the largest degree of membership, if the distances from the fan to all class centers are infinite, the degree of membership from the fan to the class centers is 0, if the fan does not belong to any class, exiting the clustering, reselecting the initial class centers, and performing clustering calculation again;
step 10: performing cluster classification on the fan, recalculating the cluster center, updating the membership degree until the cluster center is not changed any more, executing the following step 11, in the iteration process, if the fan cannot be classified into any cluster, executing the step 9, reselecting the initial cluster center, and re-clustering; if the core class of delt _ max is still changed after iteration, executing the following step 11;
and 11, step 11: and (3) obtaining a preliminary clustering result through the step 10, but the capacity of each cluster of fans is not limited in the clustering process, and the clustering result may have the overload problem, so that the clustering result needs to be subjected to overload detection, the capacity of each cluster of fans is calculated, if the capacity exceeds the double-circuit maximum limit value, the step 6 is returned, and if the capacity is not overloaded, clustering is successful, and the final subregion division result is returned.
In step 2), a cost-optimal tree is obtained for each subregion, specifically as follows:
for each sub-area, constructing a graph G (V, E), converting the line planning problem into a graph theory problem, and obtaining a cost optimal topology in each sub-area by using a mathematical method of the graph theory;
the cost optimal topology in the sub-area is obtained through the following steps:
2.1 No intersection in the construction area and the shortest path of the effective connected graph;
2.2 Searching an optimal three-dimensional path between effective paths of the fans, and taking a three-dimensional distance between the fans as a weight of an edge;
2.3 Forming a connection topological graph taking the shortest cable length as a target in a subarea;
2.4 The length of the obtained topological cable is minimum, the investment cost of the cable is not the lowest, and the result still needs to be further improved, so that an improved algorithm of dynamic adjustment and repeated optimization is provided, the problem that the section of a lead cannot be selected according to the trend distribution before planning in practice is solved, and an optimized planning topological structure with the lowest total sub-area cost is obtained through a dynamic tree algorithm; the cost of the cable between the two fans is f x l, f is the unit price cost of the cable per kilometer, the unit is ten thousand/km, and is related to the current-carrying capacity (namely the cross section of the cable), the more fans are carried by the cable, the thicker the cross section of the selected cable is, and the higher the cost of the cable is; l is the length of the cable, and the unit is km; so a minimum cable length Σ l does not represent an optimal cost Σ f × l;
2.5 Repeating the steps 2.1) to 2.4) to obtain the optimal topology of the cost of each sub-area.
In step 2.1), the method for constructing the effective connectivity map in the sub-region is as follows:
for the wind farm current collecting lines, the current collecting lines cannot be crossed, and in order to obtain a connectivity graph of the wind farm, the connectivity graph is obtained by comparing the following 3 rules:
a. the method is a full-path connection diagram, namely all fans are connected in pairs, a large number of crossed paths exist in the connection diagram, and a large number of unreasonable paths are also connected, so that the later planning is very complicated, and the path connection mode is not considered;
b. the feasible paths obtained in such a way can simplify the connection diagram, but still have the problem of intersection and filter part of the feasible paths;
c. the method adopts a feasible path obtained by Delaunay triangulation, the problem of crossing is avoided due to the connection mode, and the planned path is reasonable, so that the Delaunay triangulation is adopted to obtain a connected graph, and divides a convex polygon formed by a point set into a series of triangles, so that the condition that all edges are not crossed is ensured;
and obtaining a reasonable fan effective path communicating graph through Delaunnay triangular division.
In step 2.2), searching for an optimal three-dimensional path among the effective paths of the fans, namely searching for an optimal three-dimensional path of the effective path of the connected graph, specifically as follows:
the distance between fans in a plain area can be calculated according to a straight-line distance, but the straight-line distance between two points cannot be directly used due to the fact that a mountain area or a terrain is complex, a feasible path needs to be obtained in order to obtain a more accurate cost model, namely a three-dimensional path between the fans or between a booster station and the fans, the three-dimensional distance value is used as a side weight, an algorithm for obtaining the three-dimensional path comprises an ant colony algorithm and an RRT (rapid search tree) algorithm, wherein the ant colony algorithm is long in planning time, the RRT algorithm has the problem that the path cannot be planned, the Astar algorithm is used as one of heuristic search algorithms, and the Algorithm is an algorithm which is provided with a plurality of nodes on a graph plane and is used for solving the lowest passing cost;
therefore, the Astar algorithm is adopted to obtain the three-dimensional distance value of the feasible path, the topographic data of the wind power plant fan is firstly read, then the topographic data is subjected to grid division, each grid point has corresponding X, Y and Z values, then the Astar algorithm can be used to obtain the three-dimensional distance, and in the calculation process of the Astar algorithm, when peripheral nodes of a certain node are explored, a gradient coefficient and the terrain can be introduced to enable cost calculation to be more accurate; for a wind power plant with a complex terrain, the distance between fans cannot be simply calculated by using the linear distance between two points, the terrain and gradient problems of the wind power plant need to be considered, and a more accurate three-dimensional distance can be obtained by adopting an Astar algorithm.
In step 2.3), a connection topological graph with the shortest cable length as a target in the sub-area is formed, specifically as follows:
by the steps 2.1) and 2.2), a connection graph of the scattered points of the wind power plant and distance weights of all the connection paths are obtained, the scattered points of the fans are used as points, and the distance between the fans is used as a weight of a side, so that a graph G (V, E) can be constructed;
the topological connection optimization problem of the wind power plant current collection system can be expressed as follows: the wind power plant comprises a booster station and a plurality of wind power generators, the positions of the booster station and the wind power generators are selected, a topological connection scheme which meets the actual engineering requirements and has the best economical efficiency for a current collection system of the wind power plant needs to be obtained, and the topological connection optimization problem is a mathematical optimization problem;
the optimal cost target in the sub-region is the problem of searching a graph G (V, E) meeting known limiting conditions in a mathematical graph theory, wherein V in the graph G (V, E) is a set of all vertexes of the graph G, and represents the positions of a wind generating set and a booster station of a wind power plant; e is a set of all edges of the graph G, and represents the connection condition of cables between the booster station and the wind driven generator and between the wind driven generator and the wind driven generator; meanwhile, the graph G (V, E) is a weighted graph, and the weight of each edge is the three-dimensional distance weight between the fans represented by the edge;
the Prim algorithm is a minimum spanning tree algorithm used for solving a weighted connected graph, in a tree formed by edge subsets searched by the Prim minimum tree algorithm, not only all vertexes in the connected graph are included, but also the sum of weights of all edges is minimum; therefore, the Prim algorithm is adopted to obtain the minimum tree with the shortest distance as the target, the minimum tree comprises all fan nodes, and no ring is formed.
In step 2.4), the topological result of the lowest planning of the total sub-area cost is obtained, which specifically comprises the following steps:
setting the weight of the edge of the Prim minimum tree as a distance, setting the obtained minimum tree as a minimum tree with the minimum cable length and not a minimum tree with the minimum cable total cost, wherein the result still needs to be further improved;
the basic principle of the improved algorithm is as follows:
in a G (V, E) connectivity graph, vertexes V = { A, B, C, D, E, F, G }, edges E = { AB, BC, CD, DE, EF, FA, AG, BG, CG, EG, FG, CE }, the weight of the edges is the distance weight between two points = {12,10,3,4,8,14,16,7,6,2,9,5}, then according to a Prim algorithm, a Prim minimum tree is obtained, wherein the minimum tree comprises all vertexes, no ring exists, and the distance weight of the whole tree is minimum; calculating the cost of the Prim minimum tree as W (k);
performing dynamic adjustment on the basis of the Prim minimum tree, taking out all edges outside all the trees { BC, CG, CE, FG, AG and AF }, sequencing according to the sequence of weights from small to large to form an edge queue QV = { CE, CG, FG, BC, AF and AG }, taking out the head CE of the QV, and adding the head CE of the QV into the tree;
forming a ring, wherein the ring = { CE, ED and CD }, taking out the rest edges in the ring { CD, ED }, sequencing the edges in the ring from small to large to form a queue QE = { CD, ED }, deleting the CD edge at the head of the QE queue to form a new tree; calculating the total cost of the tree, recording as W (n), comparing the cost of the original tree with the cost of the two adjusted trees, updating the tree if the cost is better, continuing to check other edges in QE until QE is empty if the cost is not reduced, and then checking the other edges in QV until QV is empty;
the cost optimal calculation method in the sub-region comprises the following steps:
step 1: solving a minimum spanning tree by adopting a Prim algorithm;
step 2: and calculating the cost of the current T (k) tree, wherein the cost calculation method comprises the following steps:
firstly, acquiring the current-carrying capacity of each node cable according to a connection mode in a subregion, inquiring a corresponding cost coefficient according to the current-carrying capacity, accumulating and finally calculating the whole-field cost;
the cost calculation method of the m-th collecting line is as follows:
f i on the current collecting line, the cost coefficient of the ith line depends on the current carrying capacity of the cable and is in units of ten thousand/km;
l i -the length of the ith line on this line is km;
node-the number of lines on this collector line;
Cost m the cost of this current collection line;
the total cost of the wind power plant:
Cost m the cost of the mth collecting line;
n is the number of the current collecting lines, and the double-circuit line is regarded as one current collecting line;
f c the cost of the switch cabinet is ten thousand per screen;
c is the total number of switch cabinets;
a is the total cost of current collecting lines of the wind power plant;
and 3, step 3: taking out all edges except T (k), and putting the edges into a QV queue according to the sequence of the weights from small to large;
and 4, step 4: taking out the queue head from the queue QV and adding the queue head into a T (k) tree, and putting other edges in a loop caused by the addition into a queue QE according to the sequence of the weight values from small to large;
and 5, step 5: taking out the head of the queue QE and deleting the head of the queue from the loop, thereby forming a new tree T (n);
and 6, a step of: and (3) calculating the total cost of the tree T (n) according to the step 2 to be recorded as W (n), if W (n) < W (k), the new tree is more excellent in cost, and the tree is updated to be the more excellent tree in cost:
W(k)＝W(n),T(k)＝T(n)
then emptying the queue QV and QE and returning to the step 3;
if W (n) is more than or equal to W (k), judging whether the queue QE is empty, if not, returning to the step 5, if yes, judging whether the QV is empty, if so, finishing planning and exiting, wherein T (k) is an optimal planning result, the corresponding minimum total cost is W (k), and if not, returning to the step 4;
in step 3), searching feasible paths among the sub-regions, specifically as follows:
for fuzzy clustering, a certain fan may belong to one of these classes, or may belong to another class, but the probabilities of belonging to different clusters are different, and the following possibilities exist in the actual planning process: after the fan points in the area are divided into other sub-areas, the cost of topological connection is lower, and therefore feasible paths among strings are explored;
step 1: firstly, obtaining a feasible path connected graph of a whole field by adopting a Delaunay algorithm;
step 2: filtering paths inside the Delaunay connected graph sub-region and crossed paths, wherein the residual paths are feasible paths among the strings, and adding the feasible paths in the sub-region into the cost optimal tree to obtain a new connected graph;
and 3, step 3: and calling a dynamic minimum tree algorithm to obtain a final full-field cost optimal tree, wherein if the tree is overloaded, the tree cannot be used as the optimal tree.
In step 4), the T connection optimization independently performs the following processing on each series of power collecting lines, and the steps are as follows:
step 1: and acquiring the fan farthest away from the booster station in the string, and inquiring a main line on the basis of the node until the booster station:
step 2: processing the main line, starting from the booster station, detecting whether the included angle between the node and the previous node and the next node is an acute angle, if so, making a vertical line, and if so, continuing to inquire until the last fan is detected;
and 3, step 3: processing the branch, detecting whether the included angle between the branch and the main line is an acute angle, if so, making a perpendicular line, otherwise, continuously inquiring until the branch is inquired completely;
thus, a T-junction path is obtained.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the method, multiple algorithms are combined, the calculation process is reasonable, the method accords with the idea of manual design, large-scale and large-quantity planning problems are subjected to regional division and hierarchical breaking, the problems of high strength, nonlinearity and high time complexity of current collection line planning are solved, a user only needs to input topographic data of a wind power plant to be planned, and a booster station and a fan point position to complete automatic planning of a current collection line of the wind power plant, and a test case shows that the method can complete automatic design of the current collection line finally, and the obtained current collection line is low in cost, short in operation time and reliable in operation result.
In a word, the method can obtain the connection mode of the current collecting line of the wind power plant within minutes only by inputting the topographic data of the wind power plant and the point positions of the fans; in addition, the planning of the wind power plant can be conducted in a considerable and gradual mode, the influence of human factors on the planning result is reduced, the planning time of the current collecting line can be greatly shortened, and a large amount of time cost and labor cost are saved.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention.
Fig. 2 is a schematic diagram of the distance from the fan to the centroid.
Fig. 3 is a flow chart of the improved fuzzy clustering algorithm.
Fig. 4a is a full path communication diagram.
Fig. 4b is a communication diagram connecting the nearest n fans.
Fig. 4c is a Delaunay connectivity diagram.
FIG. 5a is a G (V, E) communication diagram.
FIG. 5b is a Prim minimum tree diagram.
FIG. 5c is a diagram of adding a QV head of line.
FIG. 5d is a diagram of the delete QE queue head.
FIG. 6 is a flow chart of a dynamic minimum tree obtaining cost-optimized tree optimization.
Fig. 7a is a schematic view of the connection between two fans.
Fig. 7b is a schematic diagram of a T-connection mode between fans.
FIG. 8a is a schematic view of a wind farm area for a case (reading machine location information and geographic information).
FIG. 8b is a schematic diagram of radial clustering for fuzzy clustering.
Fig. 8c is a connected graph obtained by the Delaunay triangulation algorithm.
FIG. 8d is a graph showing the results of the Astar algorithm calculating the three-dimensional path and path length.
FIG. 8e is a schematic diagram of Prim algorithm obtaining distance minimum tree.
FIG. 8f is a diagram of the dynamic minimum tree algorithm to obtain cost-optimized trees within a group.
FIG. 8g is a topological diagram of cost optimization in each sub-region of the full field.
Fig. 8h is a full-field triangulation topological graph.
FIG. 8i is a diagram of feasible paths between search sub-regions.
FIG. 8j is a diagram illustrating a dynamic minimum tree algorithm being invoked to obtain a full-field cost optimal tree.
Fig. 8k is a schematic diagram of retrieving the T-junction and performing T-junction optimization.
Detailed Description
The present invention is further illustrated by the following examples.
As shown in fig. 1, the method for automatically planning the collecting line of the multi-dimensional constraint wind farm provided by this embodiment includes the following steps:
1) Intelligent partitioning
The method comprises the steps of carrying out regional division, dividing a fan array of a wind power plant into a plurality of radioactive regions by taking a booster station as a center, limiting the capacity of each region, planning a current collecting line of the wind power plant, and carrying out grouping order reduction on the fans of the whole plant in order to reduce the planning complexity. Therefore, the method not only accords with the manual design rule, but also has the function of reducing the dimension, and replaces the manual work with the computer classification to carry out the complicated classification work. To achieve this, we use an improved fuzzy clustering algorithm.
2) Cost-optimized planning in a region
And (3) constructing a graph G (V, E) for each sub-region, converting the line planning problem into a graph theory problem, and obtaining the cost optimal topology in each sub-region by using a mathematical method of the graph theory.
3) Line cross-regional planning global cost optimization
And increasing feasible paths among the sub-regions for exploration, adding the feasible paths among the regions into a sub-region cost optimal tree, and performing dynamic optimization to obtain a final full-field cost optimal topology.
4) T-connection path optimization
The problem of T-connection intersection points is not considered in the obtained cost-optimal topology, but a certain T-connection mode exists in actual engineering, so that the obtained line topology needs to be subjected to T-connection optimization to obtain an economical and better current collection line connection topology.
In the step 1), intelligent region division is carried out by adopting a fuzzy clustering algorithm, and the specific conditions are as follows:
clustering analysis is to classify individuals or objects into categories according to similarity (or distance) so that the similarity between elements in the same category is stronger than that of other categories. The goal is to maximize the homogeneity of inter-class elements and the heterogeneity of inter-class elements. The main basis is that samples gathered in the same data set should be similar to each other, but samples belonging to different groups should be sufficiently dissimilar, and the essence is that the complex classification work is performed by replacing manual classification with computer classification. According to the method, a fuzzy clustering algorithm is adopted to divide the area of the whole field fan, and in the process of adopting the fuzzy clustering algorithm, the following two limiting conditions are required in consideration of the actual engineering requirements of the wind power plant:
a. based on the particularity of the wind power plant, radial clustering results with booster stations as the centers are required to be obtained instead of block clustering;
b. due to the limitation of the current-carrying capacity of the cable, the capacity of each current collecting line needs to be limited, and therefore capacity detection is carried out on each cluster of fans to determine whether the capacity exceeds the limit.
In the fuzzy clustering algorithm, firstly, the number of clustered clusters (namely the number of sub-regions) is determined, and the initial clustering center points have a decisive effect on the final clustering result, so that the initial clustering points are selected and distributed as uniformly as possible in consideration of the actual wind power planning, so that reasonable region division can be obtained conveniently and quickly, and therefore, the region division is performed after the traditional fuzzy clustering algorithm is improved.
Wherein the number of sub-regions is determined as follows:
for a wind farm, the number of sub-regions of the forward collector line is not determined in a connection manner, but generally, the number of divided minimum sub-regions is easily calculated according to the capacity of the wind farm and the maximum capacity limit of each sub-region. Therefore, the number of the sub-regions is clustered according to the number of the sub-regions with the minimum value, a reasonable iteration stop condition is set, and if the clustering cannot be completed, the number of the sub-regions is gradually increased until the clustering is successful.
As shown in fig. 3, the improved fuzzy clustering algorithm has the following steps:
step 1: and setting related parameters including the minimum sub-region number min _ group, the maximum sub-region number max _ group, the initial class-core maximum cycle number fuzzy _ max, the maximum cycle number station _ max of the fan not belonging to any class, the cluster-core movement maximum iteration number delt _ max and the tolerance tolerence.
Step 2: the exploration is started from the condition that the sub-area number group is the minimum set sub-area number min _ group, the cluster center number is the group number at this time, each sub-area is divided into one cluster, and the collecting lines planned and formed in each sub-area are called a string or one collecting line.
And 3, step 3: and judging whether the sub-region number group exceeds the maximum value max _ group, if not, executing the following step 4, otherwise, indicating that the clustering fails.
And 4, step 4: the method comprises the steps of dividing a wind field extension into group sub-regions, regarding a double loop as a sub-region, detecting load flow in the sub-regions after clustering is finished because the number of fans in the class, namely the capacity of each sub-region is not limited in the clustering process, and clustering again when the capacity limit is exceeded, so that a fuzzy cycle is embedded for re-circulating clustering under the condition, the fuzzy is initialized to be 0, and the maximum iteration number is fuzzy _ max.
And 5, step 5: and judging whether fuzzy is smaller than fuzzy _ max, if so, executing the following step 6, otherwise, increasing the number of planning sub-areas by group = group +1, updating, wherein the group is the number of the sub-areas, and executing the step 3.
And 6, step 6: because the distance from the fan to the class center adopts a polar coordinate form, when the included angle between the vector from the booster station to the fan and the vector from the booster station to the class center is more than 90 degrees, the fan does not belong to the class, if the fan does not belong to any class, the initial class center is reselected for clustering, so that a station cycle is embedded for re-clustering under the condition, the station is initialized to be 0, and the maximum iteration number is station _ max.
And 7, step 7: and judging whether the station is smaller than station _ max, if so, executing the next step 8, otherwise, circularly accumulating fuzzy = fuzzy +1 to update the cycle times, and executing the step 5.
And 8, step 8: the initial class center of the cluster is obtained by using a roulette algorithm, and in the selection process of the initial class center, group points which are as far away from each other as possible are selected in order to obtain a reasonable clustering effect, so that the iteration steps of the algorithm can be reduced, and the classification result is more uniform; the method mainly comprises the following steps: firstly, randomly selecting a fan point as a first initial cluster center point, then selecting the point farthest from the fan point as a second initial cluster center point, then selecting the point farthest from the first two points as a third initial cluster center point, and so on until group initial cluster center points are selected; after the initial center of class is determined, that is, all center points of the center of class are completely determined, the clustering result is determined, so that the clustering result lacks diversity and even clustering failure can be caused.
Step 9: calculating the distance d from each fan point to each class center ic I belongs to (1, n _node), c belongs to (1, n _c), i represents the ith fan, n _ node represents the number of fans, c represents the class center, n _ c represents the number of class centers, d ic Expressing the distance from the ith fan to the c class center, carrying out normalization processing on the distance, converting the distance into a membership degree matrix from each fan to each class center, and dividing fan nodes into different clusters according to the membership degree; if the fan does not belong to any class, namely the distances from the fan nodes to all the class centers are infinite, the initially selected class centers cannot finish clustering, the initially selected class centers need to be reselected, and the station + + (station + + represents that one is accumulated in each cycle of station and is used for counting) returns to the 7 th step, the initially selected class centers are reselected for calculation, otherwise, the next 10 th step is continuously executed; the method for calculating the distance from the fan node to the class center comprises the following steps: in order to obtain radial clustering results, the vertical distance from the fan to the booster station and the center of the class is taken as the distance from the fan to the center of the class instead of the straight distance from the fan to the center of the class, which is shown in fig. 2.
When the booster station arrives at the class center phasor
Is equal to the amount from the booster station to the fan>
When the included angle is less than or equal to 90 degrees, the distance from the fan to the class center is the phase quantity from the fan>
Perpendicular distance d of ic = d, if the angle is greater than 90 °, then>
And &>
Is greater than 90, the distance from the fan to the class center is set as d jc ＝∞；
The membership of the fuzzy clusters is calculated as follows:
d ic -is the distance from the ith fan to the c-th centroid;
n _ c-represents the number of centroids;
m-is a weighted index;
member ic the membership degree from the ith fan to the c class center;
for a certain fan, dividing fan nodes into clusters with the largest membership degree, if the distances from the fan to all class centers are positive and infinite, the membership degree values from the fan to the class centers are all 0, and if the fan does not belong to any class, exiting the clustering, reselecting the initial class centers, and performing clustering calculation again.
Step 10: performing cluster classification on the fan, recalculating the cluster center, updating the membership degree until the cluster center is not changed any more, executing the following step 11, in the iteration process, if the fan cannot be classified into any cluster, executing the step 9, reselecting the initial cluster center, and re-clustering; if the deltamax centroid still changes after iteration, step 11 below is performed.
And 11, a step of: and (4) obtaining a preliminary clustering result through the step 10, but the capacity of each cluster of fans is not limited in the clustering process, and the clustering result possibly has the overload problem, so that the clustering result needs to be subjected to overload detection, the capacity of each cluster of fans is calculated, if the capacity exceeds the double-circuit maximum limit value, the step 6 is returned, and if the capacity is not overloaded, clustering is successful, and the final subregion division result is returned.
In step 2), a cost-optimal tree is obtained for each subregion, specifically as follows:
and constructing a graph G (V, E) for each subregion, converting the line planning problem into a graph theory problem, and obtaining the cost optimal topology in each subregion by using a mathematical method of the graph theory.
The optimal cost topology in the sub-area is obtained through the following steps:
2.1 No intersection in the construction area and the shortest path of the effective connected graph;
2.2 Searching an optimal three-dimensional path between effective paths of the fans, and taking a three-dimensional distance between the fans as a weight of an edge;
2.3 Forming a connection topological graph taking the shortest cable length as a target in a subarea;
2.4 The length of the obtained topological cable is minimum, the investment cost of the cable is not the lowest, and the result still needs to be further improved, so that an improved algorithm of dynamic adjustment and repeated optimization is provided, the problem that the section of a lead cannot be selected according to the power flow distribution before planning in practice is solved, and the optimized planning topological structure with the lowest total sub-area cost is obtained through a dynamic tree algorithm. The cost of the cable between the two fans is f x l, f is the unit price cost of the cable per kilometer, the unit is ten thousand/km, and is related to the current-carrying capacity (namely the cross section of the cable), the more fans are carried by the cable, the thicker the cross section of the selected cable is, and the higher the cost of the cable is; l is the length of the cable, and the unit is km; so a minimum cable length Σ l does not represent an optimal cost Σ f × l.
2.5 Repeat the above steps 2.1) -2.4) to obtain a cost-optimal topology for each sub-region.
In step 2.1), the method for constructing the effective connectivity graph in the sub-region is as follows:
for the wind farm collecting lines, the collecting lines cannot be crossed, and in order to obtain the connection graph of the wind farm, as shown in fig. 4a to 4c, the connection graph is obtained by comparing the following 3 rules:
a. the method is a full-path connection diagram, namely all fans are connected in pairs, a large number of crossed paths exist in the connection diagram, and a large number of unreasonable paths are also connected, so that the later planning is very complicated, and the path connection mode is not considered;
b. the method adopts n fans nearest to a certain fan as feasible paths, and the feasible paths obtained in the way can simplify a communication graph, but still have the problem of intersection, and filter out part of feasible paths;
c. the connection mode has no crossing problem, and the planned path is reasonable, so that the Delaunay triangulation is selected to obtain a connected graph, and divides a convex polygon formed by a point set into a series of triangles, so that no crossing exists between all edges;
through Delaunnay triangular division, a reasonable fan effective path communicating graph can be obtained.
In step 2.2), searching for an optimal three-dimensional path among the effective paths of the fans, namely searching for an optimal three-dimensional path of the effective path of the connected graph, specifically as follows:
the distance between fans in a plain area can be calculated according to a linear distance, but the linear distance between two points cannot be directly used due to the fact that a mountain area or a terrain is complex, in order to obtain a more accurate cost model, a feasible path (a three-dimensional path between the fans or between a booster station and the fans) needs to be obtained, the three-dimensional distance value is used as a side weight, an ant colony algorithm and an RRT (rapid search tree) algorithm are obtained, the ant colony algorithm is longer in planning time, the RRT algorithm has the problem that the path cannot be planned, the Astar algorithm is used as one of heuristic search algorithms, and the Algorithm is an algorithm which is used for solving the lowest passing cost by the path with multiple nodes on a graphic plane.
Therefore, the Astar algorithm is adopted to obtain the three-dimensional distance value of the feasible path, the topographic data of the wind power plant fan is firstly read, then the topographic data is subjected to grid division, each grid point has corresponding X, Y and Z values, then the Astar algorithm can be used to obtain the three-dimensional distance, in the calculation process of the Astar algorithm, when the peripheral nodes of a certain node are explored, gradient coefficients can be introduced, and the cost calculation is more accurate due to the terrain (such as different terrain lawns, whether land is required to be used, and the like). For a wind power plant with a complex terrain, the distance between fans cannot be simply calculated by using the linear distance between two points, the terrain, the gradient and other problems of the wind power plant need to be considered, and a more accurate three-dimensional distance can be obtained by adopting an Astar algorithm.
In step 2.3), a connection topological graph with the shortest cable length as a target in the sub-area is formed, which specifically includes the following steps:
through steps 2.1) and 2.2), we have obtained the connection graph of the wind farm scatter points and the distance weights of the communication paths, we take the fan scatter points as points, and take the distance between the fans as the weight of the edges, so that graph G (V, E) can be constructed.
The topological connection optimization problem of the wind power plant current collection system can be expressed as follows: the wind power plant comprises a booster station and a plurality of wind power generators, the positions of the booster station and the wind power generators are selected, a topological connection scheme which meets the actual engineering requirements and has the best economical efficiency for a wind power plant current collection system needs to be obtained, and the topological connection optimization problem is a mathematical optimization problem.
The optimal cost target in the sub-region is the problem of searching a graph G (V, E) meeting known limiting conditions in a mathematical graph theory, wherein V in G (V, E) is a set of all vertexes of the graph G and represents the positions of a wind generating set and a booster station of a wind power plant in the algorithm; and E is a set of all edges of the graph G, and shows the cable connections between the booster station and the wind turbine generator and between the wind turbine generator and the wind turbine generator. Meanwhile, the graph G (V, E) is a weighted graph, and the weight of each edge is the three-dimensional distance weight between the fans represented by the edge.
The Prim algorithm is a minimum spanning tree algorithm used for solving a weighted connected graph, in a tree formed by edge subsets searched by the minimum tree algorithm, not only all vertexes in the connected graph are included, but also the sum of weights of all edges is minimum. Therefore, the Prim algorithm is adopted to obtain the minimum tree with the shortest distance as the target, the minimum tree comprises all fan nodes, and a ring is not formed.
In step 2.4), the topological result of the lowest planning of the total sub-area cost is obtained, which specifically comprises the following steps:
the weight of the edge of the Prim minimum tree is set as a distance, the obtained minimum tree is the minimum tree with the minimum cable length and is not the minimum tree with the minimum cable total cost, the result can be further improved, in order to obtain the cost optimal tree, an improved algorithm of dynamic adjustment and repeated optimization is provided, and the problem that the section of a wire cannot be selected according to the trend distribution before planning in practice is solved, namely, the new tree is formed by adding and deleting edges on the basis of the Prim minimum tree with the distance, and the cost minimum tree is obtained by detecting whether the cost of the new tree is more optimal or not.
The basic principle of the improved algorithm is as follows:
FIG. 5a is a G (V, E) connectivity graph, containing vertices V = { A, B, C, D, E, F, G }, edges E = { AB, BC, CD, DE, EF, FA, AG, BG, CG, EG, FG, CE }, where the weight of an edge is the distance weight between two points = {12,10,3,4,8,14,16,7,6,2,9,5}, according to the Prim algorithm, a minimum tree shown in FIG. 5B can be obtained, which contains all vertices, no rings exist, and the distance weight of the entire tree is minimum; the cost of calculating the Prim minimum tree is W (k).
And performing dynamic adjustment on the basis of the Prim minimum tree. All edges outside all the trees are taken out { BC, CG, CE, FG, AG, AF }, and are sorted according to the order of the weights from small to large to form an edge queue QV = { CE, CG, FG, BC, AF, AG }, and a queue head CE of the QV is taken out and added into the tree, which is shown in FIG. 5 c.
At this time, a ring is formed, the ring = { CE, ED, CD }, the remaining edges in the ring are taken out of { CD, ED }, the queue QE = { CD, ED } is formed by sorting the edges from small to large, the queue QE = { CD, ED } is formed, the CD edge at the head of the QE queue is deleted, a new tree is formed, as shown in fig. 5d, the total cost of the tree is calculated and marked as W (n), the cost of the original tree and the cost of the two adjusted trees are compared, if the cost is more excellent, the tree is updated, if the cost is not reduced, the other edges in the QE are continuously checked until the QE is empty, and the remaining edges in the QV are checked until the QV is empty.
The optimal cost calculation method in the sub-area is as follows:
step 1: solving a minimum spanning tree by adopting a Prim algorithm;
step 2: and calculating the cost of the current T (k) tree, wherein the cost calculation method comprises the following steps:
firstly, acquiring the current-carrying capacity of each node cable according to a connection mode in a subregion, inquiring a corresponding cost coefficient according to the current-carrying capacity, accumulating and finally calculating the whole-field cost;
the cost calculation method of the m-th collecting line is as follows:
f i on the current collecting line, the cost coefficient of the ith line depends on the current carrying capacity of the cable and is in units of ten thousand/km;
l i -the length of the ith line on this line is km;
node-the number of lines on this collector line;
Cost m the cost of this current collection line;
the total cost of the wind power plant is as follows:
Cost m the cost of the mth collecting line;
n-number of collecting lines (double circuit lines are regarded as a collecting line)
f c The cost of the switch cabinet is ten thousand per screen;
c is the total number of switch cabinets;
a is the total cost of current collecting lines of the wind power plant;
and 3, step 3: taking out all edges except T (k), and putting the edges into a QV queue according to the order of the weights from small to large;
and 4, step 4: taking out a queue head from the queue QV and adding the queue head into a T (k) tree, and putting other edges in a loop caused by the queue head into a queue QE according to the sequence of weight values from small to large;
and 5, step 5: taking out the head of the queue QE and deleting the head from the loop, thereby forming a new tree T (n);
and 6, step 6: and (3) calculating the total cost of the tree T (n) according to the step 2 to be recorded as W (n), if W (n) < W (k), the new tree is more excellent in cost, and the tree is updated to be the more excellent tree in cost:
W(k)＝W(n),T(k)＝T(n)
then emptying the queue QV and QE and returning to the step 3;
if W (n) is not less than W (k), judging whether the queue QE is empty, if not, returning to the step 5, if so, judging whether the QV is empty, if so, finishing planning and exiting, wherein T (k) is an optimal planning result, the corresponding minimum total cost is W (k), and if not, returning to the step 4.
The dynamic minimum tree acquisition cost-optimal tree optimization flow is shown in fig. 6.
In step 3), searching feasible paths among the sub-regions, which is specifically as follows:
for fuzzy clustering, a certain fan may belong to the same category or other categories, but the probability of belonging to different clusters is different, and there is a possibility in the actual planning process: after the fan points in the area are divided into other sub-areas, the cost is lower when topological connection is carried out, so that feasible paths among strings are explored;
step 1: firstly, obtaining a feasible path connected graph of a whole field by adopting a Delaunay algorithm;
step 2: filtering paths inside the Delaunay connected graph sub-region and crossed paths, wherein the residual paths are feasible paths among the strings, and adding the feasible paths in the sub-region into the cost optimal tree to obtain a new connected graph;
and 3, step 3: and calling a dynamic minimum tree algorithm to obtain a final full-field cost optimal tree, wherein if the tree is overloaded, the tree cannot be used as the optimal tree.
In the step 4), a topological graph of the power connection lines of the wind power plant can be obtained through the steps, the connection modes of the wind power plants are pairwise connection, but in general practical engineering, a T-shaped connection mode also exists, and the mode can save more cost, as shown in fig. 7a and 7 b.
Each series of the power collecting lines are independently processed as follows:
step 1: and acquiring the fan farthest away from the booster station in the string, and inquiring a main line on the basis of the node until the booster station:
step 2: processing a main line, starting from a booster station, detecting whether an included angle between the node and the previous node and the next node is an acute angle, if so, making a vertical line, and if so, continuing to inquire until the last fan is detected;
and 3, step 3: processing the branch, detecting whether the included angle between the branch and the main line is an acute angle, if so, making a perpendicular line, otherwise, continuously inquiring until the branch is inquired completely;
thus, a T-connection path can be obtained.
Taking a certain wind power plant as an example, the test is carried out; as shown in fig. 8a, a wind farm region and a fan point bitmap are obtained, geographical information and fan point information of the wind farm are read, wherein a 0 point is represented as a booster station point, 1-50 points represent fan points, the number of fans of a single-return current collection line is at most 6, the number of fans of a double-return current collection line is at most 12, and the operation time of the algorithm in the embodiment is 2 minutes and 05 seconds.
The fuzzy clustering grouping result is shown in fig. 8 b.
Taking a certain subregion as an example for analysis, obtaining a connected graph by using a Delaunay triangulation algorithm, and calling an Astar algorithm to calculate the distance between the fans as the weight of the side, as shown in FIGS. 8c and 8 d.
Using Prim minimum tree algorithm, the minimum tree targeting the shortest distance is obtained, as shown in fig. 8 e.
And calling a dynamic minimum tree algorithm to obtain a cost optimal tree with the cost optimal as a target, as shown in fig. 8 f.
And planning the cost optimal tree in the sub-regions for the rest sub-regions to obtain the cost optimal tree of each sub-region in the whole field, as shown in fig. 8 g.
And (5) obtaining a full-field fan connection mode by using a triangulation algorithm, as shown in fig. 8 h.
And filtering the intra-string connection and the cross path to obtain an inter-string path, and calling an Astar algorithm to obtain an inter-string three-dimensional path weight, as shown in fig. 8 i.
And calling a dynamic minimum tree algorithm to perform sub-region interaction to obtain an optimized full-field cost optimal tree, as shown in fig. 8 j.
And (5) performing T-connection retrieval, and performing T-connection processing on branches which can be subjected to T-connection to obtain a final full-field collecting circuit topological diagram, as shown in fig. 8 k.
The above-mentioned embodiments are only preferred embodiments of the present invention, and the scope of the present invention is not limited thereby, and all changes made in the shape and principle of the present invention should be covered within the scope of the present invention.
Claims (8)
1. A multidimensional constraint wind power plant current collection circuit automatic planning method is characterized by comprising the following steps:
1) Intelligent partitioning
The method comprises the following steps of performing regional division, dividing a fan array in a wind power plant into a plurality of radioactive regions by taking a booster station as a center, and limiting the capacity of each region, wherein the planning of a current collecting line of the wind power plant is a multi-dimensional and nonlinear planning problem;
2) Cost optimized planning in a region
For each subregion, constructing a graph G (V, E), converting the line planning problem into a graph theory problem, and obtaining the cost optimal topology in each subregion by using a mathematical method of the graph theory; wherein, graph G (V, E) refers to a path connection graph, V represents a point of the graph, and E represents an edge of the graph;
the cost optimal topology in the sub-area is obtained through the following steps:
2.1 No intersection in the construction area and the shortest path of the effective connected graph;
2.2 Searching for an optimal three-dimensional path among the effective paths of the fans, and taking the three-dimensional distance among the fans as a weight of an edge;
2.3 Forming a connection topological graph taking the shortest cable length as a target in a subarea;
2.4 The length of the obtained topological cable is minimum, the investment cost of the cable is not the lowest, and the result still needs to be further improved, so that an improved algorithm of dynamic adjustment and repeated optimization is provided, the problem that the section of a lead cannot be selected according to the trend distribution before planning in practice is solved, and an optimized planning topological structure with the lowest total sub-area cost is obtained through a dynamic tree algorithm; the cost of the cable between the two fans is f x l, f is the unit price cost of the cable per kilometer, the unit is ten thousand/km, and the cost is related to the current-carrying capacity of the cable, namely the cross section of the cable; l is the length of the cable, and the unit is km; so a minimum cable length Σ l does not represent an optimal cost Σ f × l;
2.5 Repeating the steps 2.1) -2.4) to obtain the optimal topology of the cost of each subarea;
3) Line cross-regional planning global cost optimization
Increasing feasible paths among the sub-regions for exploration, adding the feasible paths among the regions into a sub-region cost optimal tree, and performing dynamic optimization to obtain a final full-field cost optimal topology;
4) T-connect path optimization
The problem of T-connection intersection points is not considered in the obtained cost-optimal topology, and some T-connection modes exist in actual engineering, so that the obtained line topology needs to be subjected to T-connection optimization to obtain an economical and better current collection line connection topology.
2. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized in that: in the step 1), intelligent region division is carried out by adopting a fuzzy clustering algorithm, and the specific conditions are as follows:
clustering analysis, which is to classify individuals or objects according to similarity degree or distance so that the similarity between elements in the same class is stronger than that of elements in other classes, and aims to maximize the homogeneity of elements between classes and the heterogeneity of elements between classes, wherein the main basis is that samples gathered in the same data set are similar to each other, but samples belonging to different groups are dissimilar, and the essence is to perform complicated classification work by using computer classification instead of manual classification; the fuzzy clustering algorithm is adopted to divide the whole wind turbine into regions, and in the process of adopting the fuzzy clustering algorithm, the following two limiting conditions are required to be carried out in consideration of the actual engineering requirements of the wind power plant:
a. based on the particularity of the wind power plant, radial clustering results with booster stations as the centers are required to be obtained instead of block clustering;
b. due to the limitation of the current-carrying capacity of the cable, the capacity of each current collecting line needs to be limited, so that the capacity of each cluster of fans is detected, and whether the capacity exceeds the limit is judged;
in the fuzzy clustering algorithm, firstly, the number of clustered clusters, namely the number of sub-regions is determined, and an initial clustering center point plays a decisive role in a final clustering result, and the initial clustering points are selected according to wind power planning practice and are uniformly distributed as far as possible in a dispersing manner so as to conveniently and rapidly obtain reasonable region division, so that the region division is performed after the improvement of the traditional fuzzy clustering algorithm;
wherein the number of sub-regions is determined as follows:
for a wind power plant, the number of sub-regions of a front collector line is not determined in a connection mode, but the minimum number of the sub-regions can be obtained through calculation according to the capacity of the wind field and the maximum capacity limit of each sub-region, so that the number of the sub-regions is clustered according to the minimum number of the sub-regions, a reasonable iteration stop condition is set, and if clustering cannot be completed, the number of the sub-regions is gradually increased until clustering is successful;
the improved fuzzy clustering algorithm comprises the following steps:
step 1: setting related parameters including the minimum sub-region number min _ group, the maximum sub-region number max _ group, the initial class-core maximum cycle number fuzzy _ max, the maximum cycle number station _ max of the fan not belonging to any class, the cluster-core movement maximum iteration number delt _ max and the tolerance tolerence;
step 2: the method comprises the following steps that exploration is started when the sub-area number group is the minimum set sub-area number min _ group, the cluster centers are grouped at the moment, each sub-area is divided into a cluster, and the current collecting lines planned and formed in each sub-area are called a string or a current collecting line;
and 3, step 3: judging whether the sub-region number group exceeds the maximum value max _ group, if not, executing the following step 4, otherwise, indicating that clustering fails;
and 4, step 4: dividing the wind field extension into group sub-regions, regarding the double-loop as a sub-region, because the number of fans in the class, namely the capacity of each sub-region, is not limited in the clustering process, after the clustering is finished, the load flow in each sub-region needs to be detected, and when the capacity limit is exceeded, the clustering needs to be carried out again, so that a fuzzy cycle is embedded for the re-circulating clustering under the condition, the fuzzy is initialized to be 0, and the maximum iteration number is fuzzy _ max;
and 5, step 5: judging whether fuzzy is smaller than fuzzy _ max, if yes, executing the following step 6, otherwise, group = group +1, increasing the number of planning sub-areas and updating, wherein the group is the number of the sub-areas, and executing the step 3;
and 6, a step of: because the distance from the fan to the class center adopts a polar coordinate form, when the included angle between the vector from the booster station to the fan and the vector from the booster station to the class center is more than 90 degrees, the fan does not belong to the class, if the fan does not belong to any class, the initial class center is reselected for clustering, so that a station cycle is embedded for re-clustering under the condition, the station is initialized to be 0, and the maximum iteration number is station _ max;
and 7, step 7: judging whether the station is smaller than station _ max, if so, executing the next step 8, otherwise, circularly accumulating fuzzy = fuzzy +1 to update the cycle times, and executing the step 5;
and 8, step 8: the initial class center of the cluster is obtained by using a roulette algorithm, and in the selection process of the initial class center, group points which are as far away from each other as possible are selected in order to obtain a reasonable clustering effect, so that the iteration steps of the algorithm can be reduced, and the classification result is more uniform; the method mainly comprises the following steps: firstly, randomly selecting a fan point as a first initial cluster center point, then selecting the point farthest away from the fan point as a second initial cluster center point, then selecting the point farthest away from the first two points as a third initial cluster center point, and so on until group initial cluster center points are selected; after the initial center of class is determined, namely all center points of the center of class are completely determined, the clustering result is determined, so that the clustering result lacks diversity and even clustering failure can be caused, a roulette algorithm is introduced, the probability that the point is selected is higher as the distance is farther, and the point with the farthest distance is not selected by 100%, so that the uniform distribution of the initial point is ensured, and various clustering possible results are provided for clustering;
step 9: calculating the distance d from each fan point to each class center ic I e (1, n _node), c e (1, n _c), i represents the ith fan, n _ node represents the number of fans, c represents the class center, n _ c represents the number of class centers, d ic Expressing the distance from the ith fan to the c class center, carrying out normalization processing on the distance, converting the distance into a membership degree matrix from each fan to each class center, and dividing fan nodes into different clusters according to the membership degree; if the fan does not belong to any class, namely the distances from the fan nodes to all class centers are infinite, the initially selected class centers cannot complete clustering, the initial class centers need to be reselected, the station + + returns to the 7 th step, the initial class centers are reselected for calculation, and otherwise, the next 10 th step is continuously executed; the method for calculating the distance from the fan node to the class center comprises the following steps: in order to obtain a radial clustering result, the vertical distance from the fan to the booster station and the connection line of the class center is the distance from the fan to the class center, and the linear distance from the fan to the class center is not used; wherein, the station + + represents that station accumulates one for each cycle for counting;
when the booster station arrives at the class center phasor
Is equal to the amount from the booster station to the fan>
When the included angle is less than or equal to 90 degrees, the distance from the fan to the class center is the phase quantity from the fan>
Perpendicular distance d of ic = d, if the angle is greater than 90 °, then>
And &>
Is larger than 90 DEG, the distance from the fan to the class center is set as d jc ＝∞；
The membership of the fuzzy clusters is calculated as follows:
d ic -is the distance from the ith fan to the c-th centroid;
n _ c — represents the number of class centers;
m is a weighted index;
member ic -is the membership of the ith fan to the c class center;
for a certain fan, dividing fan nodes into clusters with the largest degree of membership, if the distances from the fan to all class centers are infinite, the degree of membership from the fan to the class centers is 0, if the fan does not belong to any class, exiting the clustering, reselecting the initial class centers, and performing clustering calculation again;
step 10: performing cluster classification on the fan, recalculating the cluster center, updating the membership degree until the cluster center is not changed any more, executing the following step 11, in the iteration process, if the fan cannot be classified into any cluster, executing the step 9, reselecting the initial cluster center, and re-clustering; if the core class of delt _ max is still changed after iteration, executing the following step 11;
and 11, step 11: and (3) obtaining a preliminary clustering result through the step 10, but the capacity of each cluster of fans is not limited in the clustering process, and the clustering result may have the overload problem, so that the clustering result needs to be subjected to overload detection, the capacity of each cluster of fans is calculated, if the capacity exceeds the double-circuit maximum limit value, the step 6 is returned, and if the capacity is not overloaded, clustering is successful, and the final subregion division result is returned.
3. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 2.1), the method for constructing the effective connectivity map in the sub-region is as follows:
for the wind farm collecting lines, the collecting lines cannot be crossed, and in order to obtain the connected graph of the wind farm, the connected graph is obtained by comparing the following 3 rules:
a. the method is a full path communication diagram, namely all fans are connected in pairs, a large number of crossed paths exist in the communication diagram, and a large number of unreasonable paths are also connected, so that the later planning is very complicated, and the path connection mode is not considered;
b. the feasible paths obtained in such a way can simplify the connection diagram, but still have the problem of intersection and filter part of the feasible paths;
c. the method adopts a feasible path obtained by Delaunay triangulation, the problem of crossing is avoided due to the connection mode, and the planned path is reasonable, so that the Delaunay triangulation is adopted to obtain a connected graph, and divides a convex polygon formed by a point set into a series of triangles, so that the condition that all edges are not crossed is ensured;
and obtaining a reasonable fan effective path communicating graph through Delaunnay triangular division.
4. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 2.2), searching for an optimal three-dimensional path among the effective paths of the fans, namely searching for an optimal three-dimensional path of the effective path of the connected graph, specifically as follows:
the distance between fans in a plain area can be calculated according to a straight-line distance, but the straight-line distance between two points cannot be directly used due to the fact that a mountain area or a terrain is complex, a feasible path needs to be obtained in order to obtain a more accurate cost model, namely a three-dimensional path between the fans or between a booster station and the fans, the three-dimensional distance value is used as a side weight, an algorithm for obtaining the three-dimensional path comprises an ant colony algorithm and an RRT (rapid search tree) algorithm, wherein the ant colony algorithm is long in planning time, the RRT algorithm has the problem that the path cannot be planned, the Astar algorithm is used as one of heuristic search algorithms, and the Algorithm is an algorithm which is provided with a plurality of nodes on a graph plane and is used for solving the lowest passing cost;
therefore, the Astar algorithm is adopted to obtain a three-dimensional distance value of a feasible path, firstly, topographic data of a fan of a wind power plant is read, then, grid division is carried out on the topographic data, each grid point has a corresponding X, Y and Z value, then, the Astar algorithm can be used for obtaining the three-dimensional distance, and in the calculation process of the Astar algorithm, when peripheral nodes of a certain node are explored, a gradient coefficient and the terrain can be introduced, so that cost calculation is more accurate; for a wind power plant with a complex terrain, the distance between fans cannot be simply calculated by using the linear distance between two points, the terrain and gradient problems of the wind power plant need to be considered, and a more accurate three-dimensional distance can be obtained by adopting an Astar algorithm.
5. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 2.3), a connection topological graph with the shortest cable length as a target in the sub-area is formed, specifically as follows:
by the steps 2.1) and 2.2), a connection graph of the scattered points of the wind power plant and distance weights of all the connection paths are obtained, the scattered points of the fans are used as points, and the distance between the fans is used as a weight of a side, so that a graph G (V, E) can be constructed;
the topological connection optimization problem of the wind power plant collection system can be expressed as follows: the wind power plant comprises a booster station and a plurality of wind power generators, the positions of the booster station and the wind power generators are selected, a topological connection scheme which meets the actual engineering requirements and has the best economical efficiency for a current collection system of the wind power plant needs to be obtained, and the topological connection optimization problem is a mathematical optimization problem;
the optimal cost target in the sub-area is the problem of finding a graph G (V, E) meeting known limiting conditions in a mathematical graph theory, wherein V in G (V, E) is the set of all vertexes of the graph G, and represents the positions of a wind generating set and a booster station of the wind power plant; e is the set of all edges of the graph G, and represents the cable connection condition between the booster station and the wind driven generator and between the wind driven generator and the wind driven generator; meanwhile, the graph G (V, E) is a weighted graph, and the weight of each edge is the three-dimensional distance weight between the fans represented by the edge;
the Prim algorithm is a minimum spanning tree algorithm used for solving a weighted connected graph, in a tree formed by edge subsets searched by the minimum tree algorithm, not only all vertexes in the connected graph are included, but also the sum of weights of all edges is minimum; therefore, the Prim algorithm is adopted to obtain the minimum tree with the shortest distance as the target, the minimum tree comprises all fan nodes, and no ring is formed.
6. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 2.4), the topological result of the lowest planning of the total sub-area cost is obtained, which specifically comprises the following steps:
setting the weight of the edge of the Prim minimum tree as a distance, setting the obtained minimum tree as a minimum tree with the minimum cable length and not a minimum tree with the minimum cable total cost, wherein the result still needs to be further improved;
the basic principle of the improved algorithm is as follows:
in a G (V, E) connectivity graph, vertex V = { A, B, C, D, E, F, G }, edge E = { AB, BC, CD, DE, EF, FA, AG, BG, CG, EG, FG, CE }, the weight of the edge is the distance weight between two points = {12,10,3,4,8,14,16,7,6,2,9,5}, then according to the Prim algorithm, a Prim minimum tree is obtained, wherein the minimum tree comprises all vertexes, no ring exists, and the distance weight of the whole tree is minimum; calculating the cost of the Prim minimum tree as W (k);
dynamically adjusting on the basis of the Prim minimum tree, taking out all edges outside all trees { BC, CG, CE, FG, AG and AF }, sequencing according to the sequence of weight values from small to large to form an edge queue QV = { CE, CG, FG, BC, AF and AG }, taking out the head CE of the QV, and adding the head CE of the QV into the tree;
forming a ring, wherein the ring = { CE, ED and CD }, taking out the rest edges in the ring { CD, ED }, sequencing the edges in the ring from small to large to form a queue QE = { CD, ED }, deleting the CD edge at the head of the QE queue to form a new tree; calculating the total cost of the tree, recording as W (n), comparing the cost of the original tree with the cost of the two adjusted trees, updating the tree if the cost is better, continuously checking other edges in QE if the cost is not reduced until the QE is empty, and then checking the other edges in QV until the QV is empty;
the optimal cost calculation method in the sub-area is as follows:
step 1: solving a minimum spanning tree by adopting a Prim algorithm;
step 2: and calculating the cost of the current T (k) tree, wherein the cost calculation method comprises the following steps:
firstly, acquiring the current-carrying capacity of each node cable according to a connection mode in a subregion, inquiring a corresponding cost coefficient according to the current-carrying capacity, accumulating and finally calculating the whole-field cost;
the cost calculation method of the m-th collecting line is as follows:
f i on the current collecting line, the cost coefficient of the ith line depends on the current carrying capacity of the cable and is in units of ten thousand/km;
l i -the length of the ith line on this line is km;
node — number of lines on this current collection line;
Cost m the cost of this current collection line;
the total cost of the wind power plant is as follows:
Cost m the cost of the mth collecting line;
n is the number of the current collecting lines, and the double-circuit line is regarded as one current collecting line;
f c the cost of the switch cabinet is ten thousand per screen;
c is the total number of switch cabinets;
a is the total cost of current collecting lines of the wind power plant;
and 3, step 3: taking out all edges except T (k), and putting the edges into a QV queue according to the order of the weights from small to large;
and 4, step 4: taking out a queue head from the queue QV and adding the queue head into a T (k) tree, and putting other edges in a loop caused by the queue head into a queue QE according to the sequence of weight values from small to large;
and 5, step 5: taking out the head of the queue QE and deleting the head of the queue from the loop, thereby forming a new tree T (n);
and 6, step 6: and (3) calculating the total cost of the tree T (n) according to the step 2 to be recorded as W (n), if W (n) < W (k), the new tree is more excellent in cost, and the tree is updated to be the more excellent tree in cost:
W(k)＝W(n),T(k)＝T(n)
then emptying the queue QV and QE and returning to the step 3;
if W (n) is more than or equal to W (k), judging whether the queue QE is empty, if not, returning to the step 5, if yes, judging whether the QV is empty, if so, finishing planning and exiting, T (k) is an optimal planning result, the corresponding minimum total cost is W (k), and if not, returning to the step 4.
7. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 3), searching feasible paths among the sub-regions, specifically as follows:
for fuzzy clustering, a certain fan may belong to the same category or other categories, but the probability of belonging to different clusters is different, and the following possibilities exist in the actual planning process: after the fan points in the area are divided into other sub-areas, the cost is lower by performing topological connection, and feasible paths are explored according to the lower cost;
step 1: firstly, obtaining a feasible path connected graph of a whole field by adopting a Delaunay algorithm;
step 2: filtering paths inside the Delaunay connected graph sub-region and paths with crossed paths, wherein the remaining paths are feasible paths among the strings, and adding the feasible paths in the sub-region into the cost optimal tree to obtain a new connected graph;
and 3, step 3: and calling a dynamic minimum tree algorithm to obtain a final full-field cost optimal tree, wherein if the tree is overloaded, the tree cannot be used as the optimal tree.
8. The method for automatically planning the collecting circuit of the multi-dimensional constraint wind power plant according to claim 1, characterized by comprising the following steps: in step 4), the T connection optimization independently performs the following processing on each series of the power collection lines, and the steps are as follows:
step 1: and acquiring the fan farthest away from the booster station in the string, and inquiring a main line on the basis of the node until the booster station:
step 2: processing the main line, starting from the booster station, detecting whether the included angle between the node and the previous node and the next node is an acute angle, if so, making a vertical line, and if so, continuing to inquire until the last fan is detected;
and 3, step 3: processing the branch, detecting whether an included angle between the branch and the main line is an acute angle, if so, making a perpendicular line, otherwise, continuously inquiring until the branch is inquired;
thus, a T-junction path is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910433214.0A CN110334391B (en) | 2019-05-23 | 2019-05-23 | Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910433214.0A CN110334391B (en) | 2019-05-23 | 2019-05-23 | Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110334391A CN110334391A (en) | 2019-10-15 |
| CN110334391B true CN110334391B (en) | 2023-03-28 |
Family
ID=68139713
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910433214.0A Active CN110334391B (en) | 2019-05-23 | 2019-05-23 | Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110334391B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111313406B (en) * | 2020-03-02 | 2022-05-24 | 阳光新能源开发股份有限公司 | Current collecting line determining method and device and power station |
| CN113408774A (en) * | 2020-03-17 | 2021-09-17 | 北京京东振世信息技术有限公司 | Route planning method and device, storage medium and electronic equipment |
| CN111724273B (en) * | 2020-05-09 | 2023-04-14 | 中国大唐集团科学技术研究院有限公司火力发电技术研究院 | Method for automatically planning current collection circuit of offshore wind farm by adopting large-capacity wind turbine generator |
| WO2021253291A1 (en) * | 2020-06-17 | 2021-12-23 | 上海电气风电集团股份有限公司 | Wind farm layout optimization method and optimization system, and computer-readable storage medium |
| CN112052544B (en) * | 2020-08-13 | 2023-09-29 | 明阳智慧能源集团股份公司 | Wind farm collecting network design method, system, storage medium and computing equipment |
| CN112393728B (en) * | 2020-10-23 | 2022-05-31 | 浙江工业大学 | Mobile robot path planning method based on A-algorithm and RRT-algorithm |
| CN112270030B (en) * | 2020-11-09 | 2023-03-21 | 中国船舶重工集团海装风电股份有限公司 | Method, device, equipment and medium for optimal design of complex terrain wind power plant road |
| CN112990538B (en) * | 2021-02-03 | 2024-05-17 | 阳光新能源开发股份有限公司 | Method, device and equipment for determining collecting line of mountain photovoltaic power station |
| CN112836388B (en) * | 2021-03-02 | 2023-02-07 | 中国电建集团中南勘测设计研究院有限公司 | Offshore wind farm current collection circuit design method |
| CN113128123B (en) * | 2021-04-23 | 2024-09-17 | 国核电力规划设计研究院有限公司 | Marine wind farm collecting cable path planning method and device and computer equipment |
| CN114491755A (en) * | 2022-01-24 | 2022-05-13 | 阳光新能源开发股份有限公司 | Mountain land photovoltaic power station road planning method, device, equipment and storage medium |
| CN115598670B (en) * | 2022-09-29 | 2024-03-05 | 西安邮电大学 | Method for monitoring and positioning interference source by combining triangular network and ant colony algorithm |
| CN115758640B (en) * | 2022-11-21 | 2023-10-27 | 上海天华建筑设计有限公司 | Routing planning-based automatic electrical connection method for connecting pipes |
| CN116562424B (en) * | 2023-03-30 | 2024-03-22 | 上海勘测设计研究院有限公司 | Position selection method and system for offshore substation, electronic equipment and storage medium |
| CN118378853A (en) * | 2024-06-21 | 2024-07-23 | 中国电建集团西北勘测设计研究院有限公司 | Wind power plant current collection circuit planning method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9165092B2 (en) * | 2012-07-31 | 2015-10-20 | International Business Machines Corporation | Wind farm layout in consideration of three-dimensional wake |
| CN106503839B (en) * | 2016-10-14 | 2021-09-14 | 上海电力学院 | Hierarchical planning method for offshore wind farm annular current collection network |
| CN107357965A (en) * | 2017-06-16 | 2017-11-17 | 北京道亨时代科技有限公司 | A kind of path planning design method of wind power plant collection electric line |
| CN107506854B (en) * | 2017-08-04 | 2021-04-20 | 国网浙江省电力公司经济技术研究院 | 220kV power grid structure planning method considering differentiation scene |
| CN107766991A (en) * | 2017-11-10 | 2018-03-06 | 国网浙江省电力公司电力科学研究院 | A kind of plan optimization method and system of distributed power source access power distribution network |
| CN107909222B (en) * | 2017-12-11 | 2021-02-19 | 北京天润新能投资有限公司 | Wind power plant current collection line three-dimensional path planning method and system based on intelligent fan grouping |
| CN108052731B (en) * | 2017-12-11 | 2021-01-26 | 北京天润新能投资有限公司 | Method and system for planning current collection line path of wind power plant |
-
2019
- 2019-05-23 CN CN201910433214.0A patent/CN110334391B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110334391A (en) | 2019-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110334391B (en) | Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant | |
| CN111639237B (en) | Electric power communication network risk assessment system based on clustering and association rule mining | |
| CN107909222B (en) | Wind power plant current collection line three-dimensional path planning method and system based on intelligent fan grouping | |
| CN107784380B (en) | Optimization method and optimization system for routing inspection shortest path | |
| CN109992923B (en) | Phased power transmission line path planning method based on variable resolution cost surface | |
| CN111724273B (en) | Method for automatically planning current collection circuit of offshore wind farm by adopting large-capacity wind turbine generator | |
| CN112052544B (en) | Wind farm collecting network design method, system, storage medium and computing equipment | |
| CN107368648A (en) | Municipal water supply network oriented pressure monitoring point optimal arrangement method | |
| CN107437135B (en) | Novel energy storage type selection method | |
| CN109813273A (en) | A kind of agricultural machinery repetition working area determination method based on spatial analysis | |
| CN113806947B (en) | Offshore wind farm layout processing method, device and equipment | |
| CN110046213A (en) | A kind of electric power selection method for taking path distortion correction and scissors crossing correction into account | |
| CN107391745A (en) | Extensive spatial data classification fast indexing method and device | |
| CN116203982A (en) | Unmanned aerial vehicle power patrol track planning method and system | |
| CN107818338A (en) | A kind of method and system of building group pattern-recognition towards Map Generalization | |
| CN111985690B (en) | Site selection method for booster station of wind power base | |
| Emeksiz | Multi-gen genetic programming based improved innovative model for extrapolation of wind data at high altitudes, case study: Turkey | |
| CN110097636B (en) | Site selection planning method based on visual field analysis | |
| CN111985691A (en) | Site selection method for wind power plant booster station | |
| CN116777230A (en) | Intelligent substation site selection area division method based on GIS space analysis technology | |
| CN110264010A (en) | Novel rural area electric power saturation load forecasting method | |
| CN114896425A (en) | City knowledge archive construction method and system based on city brain | |
| CN112084609B (en) | Power supply partition dividing method for power industry | |
| CN113343565A (en) | Neighborhood effect mode construction and CA simulation method and system considering spatial heterogeneity | |
| Fushimi et al. | Functional cluster extraction from large spatial networks |
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 |