CN105354643A - Risk prediction evaluation method for wind power grid integration - Google Patents
Risk prediction evaluation method for wind power grid integration Download PDFInfo
- Publication number
- CN105354643A CN105354643A CN201510824706.4A CN201510824706A CN105354643A CN 105354643 A CN105354643 A CN 105354643A CN 201510824706 A CN201510824706 A CN 201510824706A CN 105354643 A CN105354643 A CN 105354643A
- Authority
- CN
- China
- Prior art keywords
- point
- wind power
- load
- risk
- electrical network
- 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.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 27
- 230000010354 integration Effects 0.000 title claims abstract description 24
- 239000013598 vector Substances 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims description 48
- 238000012502 risk assessment Methods 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 11
- 238000011161 development Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 5
- 238000000926 separation method Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 38
- 230000005611 electricity Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 4
- 238000012512 characterization method Methods 0.000 description 4
- 230000007812 deficiency Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000011068 loading method Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000000739 chaotic effect Effects 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 230000001932 seasonal effect Effects 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010291 electrical method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000036316 preload Effects 0.000 description 1
- 238000007634 remodeling Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- 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"
-
- 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
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a risk prediction evaluation method for wind power grid integration, comprising the following steps of: collecting original data of a wind power plant to obtain an initial wind power time sequence; performing phase space reconstruction on the initial wind power time sequence to obtain a phase point vector; performing K mean clustering computation on the phase point vector; introducing a weight vector to improve criterion for distance and trend of proximal points, and thereby obtaining a proximal point set; using the proximal point set as a training set of volterra self-adaptive filter to obtain wind power plant power prediction data. The risk prediction evaluation method for wind power grid integration provided by the invention eliminates the technical defect in the prior art which ignores different influence of the time sequence of own different coordinate components of the phase point on the prediction point and is prone to introducing false proximal points, effectively avoids the false proximal points, chooses the proximal point with distance and evolutionary trend both similar to the prediction point, and is thereby improved in computation accuracy and speed.
Description
Technical field
The present invention relates to power domain, particularly a kind of risk profile appraisal procedure of wind power integration electrical network.
Background technology
Along with the development of wind generating technology, improving constantly of wind energy turbine set installed capacity, the feature such as intermittence, undulatory property of wind-powered electricity generation makes it when grid-connected, become the safety operation level of a disturbing source to electrical network to make a big impact.Therefore, accurately carry out short-term wind-electricity power prediction, and utilize the electrical network after to wind power integration that predicts the outcome to carry out dynamic security assessment, the safety operation level of the electrical network after raising wind power integration is had great importance and value.
In recent years, wind power carries out short-term forecasting aspect, Volterra sef-adapting filter is fast with its training speed, the advantages such as required sample size is little obtain the concern of numerous scholars, as " Volterramodelsandthree-layerperceptrons " (MarmarelisVZ, ZhaoX..NeuralNetworks, IEEETransactionson, 1997,8 (6): 1421-1433.).But the prediction effect of Volterra sef-adapting filter is subject to uncorrelated with future position information or contributes less phase point to affect on future position.Current research shows to adopt neighbor point as training set, prove rationally to screen neighbor point, set up Local Model, the precision that can improve model as " the local SVM prediction of spatio-temporal chaotic sequences " (Zhang Jiashu, Party building is bright, and Li Heng surpasses .. Acta Physica Sinica, 2007,56 (1): 67-77.) in a literary composition, for the selection of neighbor point, current main criterion has the methods such as Euclidean distance, vector angle, the degree of association.On the other hand, in engineer applied, the validity of model not only requires that arithmetic accuracy is high, also requires that the computing velocity of model is fast simultaneously.Use K mean algorithm effectively can improve the computing velocity of model, referring to " Short-termpredictionofwindpowerwithaclusteringapproach. " (KusiakA, LiW.RenewableEnergy, 2010,35 (10): 2362-2369.)
And in the prior art, the Volterra sef-adapting filter neighbor point that wind power carries out the employing of short-term forecasting aspect calculates, the impact that the chronological order that have ignored the different coordinate components of phase point self produces future position is different, easy introducing " pseudo-neighbor point ", causes computational accuracy and the computing velocity poor stability of algorithm.
Summary of the invention
The invention reside in the above-mentioned deficiency overcoming prior art, the risk profile appraisal procedure of the wind power integration electrical network of a kind of computational accuracy, computing velocity good stability is provided.
In order to realize foregoing invention object, the technical solution used in the present invention is:
A risk profile appraisal procedure for wind power integration electrical network, comprises the following steps:
Gather wind energy turbine set raw data, obtain initial wind power time series;
Phase space reconfiguration is carried out to described initial wind power time series, obtains phase point vector;
The calculating of K mean cluster is carried out to described phase point vector;
Introduce weight vectors and improve the distance of neighbor point and the criterion of trend, set up comprehensive criterion, obtain neighbor point set;
Using the training set of described neighbor point set as volterra sef-adapting filter, obtain wind farm power prediction data.Further, described introducing weight vectors improves the distance of neighbor point and the criterion of trend, sets up comprehensive criterion, obtains neighbor point set and comprise:
Define a kind of compute mode:
Wherein A, B are the vector of m dimension, a
i, b
ibe respectively the i-th dimension component of vectorial A, B;
Distance then between current predictive point X (p) and phase point X (i) is:
In formula, X (p) represents p future position, X (i) represents i-th phase point, α is weight vectors, and for the vectorial α of m dimension, α (1)≤α (2)≤... ≤ α (m), considers that interval time between coordinate components is for being τ, gets
Wherein, d (p, i) is less, and current predictive point X (p) is nearer with the distance of phase point X (i);
The difference value vector of definition multistep backtracking is:
E(p,q)＝X(p)-X(p-q)，
E(i,q)＝X(i)-X(i-q)，
Wherein, q represents backtracking step-length, and the q that X (p) is X (p-q) walks evolution phase point, and the q that X (i) is X (i-q) walks evolution phase point;
Angular separation between the future position of multistep backtracking and phase point is:
Be weighted above formula, the development trend criterion that can obtain future position and phase point is:
Cos θ (p, i) be by vector between folder cosine of an angle develop, wherein, β is weight vectors, and β (0)≤β (1)≤... ≤ β (q), wherein
Cos θ (p, i) is less, represents that current predictive point X (p) is more close with the development trend of phase point X (i);
Similarity between future position X (p) and phase point X (i) is:
η(p,i)＝γ
1d(p,i)+γ
2cosθ(p,i)，
Wherein γ
1, γ
2be respectively the weighted value of range index and evolution trend index, and γ
1+ γ
2=1.
To described sequencing of similarity, select the larger multiple values of similarity as neighbor point.
Further, also comprise:
Wind energy turbine set load data, obtains load prediction data according to RBF neural model;
Wind power and predicted load is obtained according to described wind farm power prediction data and load prediction data;
Selected multiple evaluation time point, carries out risk assessment according to the wind power of each described evaluation time section and predicted load to electrical network, obtains multiple risk evaluation result;
Power grid risk prediction curve is obtained according to described multiple risk evaluation result.
Further, the described wind power according to each described evaluation time section and predicted load carry out risk assessment to electrical network and comprise:
According to the importance degree of element each in electrical network, set up respectively and characterize grid nodes, the pitch point importance factor of branch road importance degree, branch road importance factors;
Introduce described branch road importance factors, revise the branch road overload severity function of network system, introduce the described pitch point importance factor, revise the node low-voltage severity function of network system;
The severity index characterizing operation of power networks state is obtained according to described branch road overload severity function, described node low-voltage severity function;
Introduce load economic factor, obtain the mistake load severity function characterizing load loss;
Respectively process is weighted to described operation of power networks severity index, mistake load severity function, obtains the comprehensive severity function characterizing power grid accident.
Further, also comprise: obtain integrated risk index according to described comprehensive severity function and contingency occurrence probability, and according to described index, risk assessment is carried out to network system.
Compared with prior art, beneficial effect of the present invention
1, the risk profile appraisal procedure of a kind of wind power integration electrical network of the present invention is when utilizing Volterra sef-adapting filter to calculate, consider the multistep evolution trend between the current distance of future position and phase point and phase point, both the evolution correlativity of phase point had been considered, consider again each coordinate of phase point Different Effects in time, eliminate in prior art, " chronological order that have ignored the different coordinate components of phase point self is different on the impact of future position, easy introducing ' pseudo-neighbor point ' " technological deficiency, therefore can effectively avoid " false neighbor point ", select at all similar to the future position neighbor point of Distance geometry evolution trend, improve computational accuracy and computing velocity.
2, after wind-powered electricity generation is connected to the grid, there is undulatory property and uncertainty, the running status of system is in the state of a change always, its security risk is also in change, and prior art does not have a good method to assess the electric network security risk after wind power integration electrical network, the risk profile appraisal procedure of a kind of wind power integration electrical network of the present invention, consider the evaluation time window of risk assessment, the value-at-risk of each time point is carried out utilization system risk walk power curve can be system cloud gray model personnel prediction and grasp system running state, can be good at predicting wind power, thus effectively prevention to be fluctuated the power system accident caused due to wind-powered electricity generation, avoid causing large-area power-cuts.
Accompanying drawing explanation
It is the wind farm power prediction process flow diagram of the risk profile appraisal procedure of a kind of wind power integration electrical network shown in a specific embodiment of the present invention shown in Fig. 1.
It is the risk profile appraisal procedure process flow diagram of a kind of wind power integration electrical network shown in a specific embodiment of the present invention shown in Fig. 2.
Embodiment
Below in conjunction with embodiment, the present invention is described in further detail.But this should be interpreted as that the scope of the above-mentioned theme of the present invention is only limitted to following embodiment, all technology realized based on content of the present invention all belong to scope of the present invention.
Embodiment 1:
Be the process flow diagram of the risk profile appraisal procedure of a kind of wind power integration electrical network shown in a specific embodiment of the present invention shown in Fig. 1, comprise the following steps:
A risk profile appraisal procedure for wind power integration electrical network, comprises the following steps:
Gather wind energy turbine set raw data, obtain initial wind power time series; Phase space reconfiguration is carried out to described initial wind power time series, obtains phase point vector.
Concrete, if initially wind power time series is x (1), x (2) ..., x (N) }, wherein N is the collection point sum of wind power, and obtaining phase point vector after phase space reconfiguration is:
As can be seen from the above equation, to sequence carry out phase space reconfiguration be at all obtain seasonal effect in time series delay time T and Embedded dimensions m, C-C method can be used to ask for.
The calculating of K mean cluster is carried out to described phase point vector.
Concrete, the basic thought of K mean cluster, for data acquisition X is divided into k class, makes the member in class there is the process of certain similarity.
Namely C={C is found
1, C
2..., C
k, meet:
And make total inter _ class relationship and
Reach minimum, wherein, M (j) is the center of a jth cluster, and d (X (i), M (j)) is for sample is to the distance of corresponding cluster centre, and its expression formula is:
d(X(i),M(j))＝||X(i)-M(j)||
2(5)
There is a great drawback in the K means clustering algorithm due to prior art, namely convergence and speed of convergence depend on starting condition.Therefore, the present invention use subtractive clustering method independently for K means Method provide initial cluster center Mj (j=1,2 ... k) with cluster number k.
The risk profile appraisal procedure of a kind of wind power integration electrical network of the present invention uses K mean cluster to process wind power data that are complicated, large sample, cluster speed is fast, Clustering Effect is good, and the stability of algorithm is strengthened, thus improves computational accuracy and the computing velocity of wind energy turbine set prediction.
Introduce weight vectors and improve the distance of neighbor point and the criterion of trend, obtain neighbor point set; Using the training set of described neighbor point set as volterra sef-adapting filter, obtain wind farm power prediction data.
Concrete, the Volterra series expansion expression of the kernel function of Volterra sef-adapting filter is infinite series form, in actual applications, usually takes limited Trimmed sums limited number of time summation form.Be below p rank Truncation:
By the desirable N of Taken embedding theorems
1=N
2=...=N
p=m,
for predicted value, h
p(m
1, m
2..., m
p) being called p rank Volterra core, m is the input dimension of wave filter, corresponds to the Embedded dimensions of wind-powered electricity generation temporal power.
Define a kind of new compute mode:
Wherein A, B are the vector of m dimension, a
i, b
ibe respectively the i-th dimension component of vectorial A, B;
Distance then between current predictive point X (p) and phase point X (i) is:
In formula, X (p) represents p future position, X (i) represents i-th phase point, α is weight vectors, and for the vectorial α of m dimension, α (1)≤α (2)≤... ≤ α (m), considers that interval time between coordinate components is for being τ, gets
Wherein, d (p, i) is less, and current predictive point X (p) is nearer with the distance of phase point X (i);
The difference value vector of definition multistep backtracking is:
E(p,q)＝X(p)-X(p-q)，(9)
E(i,q)＝X(i)-X(i-q)，(10)
Wherein, q represents backtracking step-length, and the q that X (p) is X (p-q) walks evolution phase point, and the q that X (i) is X (i-q) walks evolution phase point;
Angular separation between the future position of multistep backtracking and phase point is:
Be weighted above formula, the development trend criterion that can obtain future position and phase point is:
Cos θ (p, i) be by vector between folder cosine of an angle develop, wherein, β is weight vectors, and β (0)≤β (1)≤... ≤ β (q) gets herein
Cos θ (p, i) is less, and current predictive point X (p) is more close with the development trend of phase point X (i);
Similarity between future position X (p) and phase point X (i) is:
η(p,i)＝γ
1d(p,i)+γ
2cosθ(p,i)，(13)
Wherein γ
1, γ
2be respectively the weighted value of range index and evolution trend index, and γ
1+ γ
2=1;
To described sequencing of similarity, select the larger multiple values of similarity as neighbor point.
According to Hannan-Quinn criterion, described neighbor point set is screened, the neighbor point set be improved.
Concrete, the neighbor point number of screening is:
In formula, x
jfor the sample point of data,
for predicting the outcome,
for sample point average, S is prediction step number, and N is fitting data number, and when Φ (K) obtains minimum value, corresponding K is the number of best neighbor point.
The risk profile appraisal procedure of a kind of wind power integration electrical network of the present invention is when utilizing Volterra sef-adapting filter to calculate, consider the multistep evolution trend between the current distance of future position and phase point and phase point, both the evolution correlativity of phase point had been considered, consider again each coordinate of phase point Different Effects in time, eliminate in prior art, " chronological order that have ignored the different coordinate components of phase point self is different on the impact of future position, easy introducing ' pseudo-neighbor point ' " technological deficiency, therefore can effectively avoid " false neighbor point ", select at all similar to the future position neighbor point of Distance geometry evolution trend, further improve computational accuracy and computing velocity.
Further, also comprise: wind energy turbine set load data, obtain load prediction data according to RBF neural model;
Wind power and predicted load is obtained according to described wind farm power prediction data and load prediction data;
Selected multiple evaluation time point, carries out risk assessment according to the wind power of each described evaluation time section and predicted load to electrical network, obtains multiple risk evaluation result;
Power grid risk prediction curve is obtained according to described multiple risk evaluation result.
RBF neural structure is simple, training is succinct, pace of learning is fast, can effectively avoid local minimum and realize global convergence, can also approach complicated nonlinear function, be therefore widely used in electrical methods load prediction with arbitrary accuracy.
The neuron models of radial basis function neural network, it is a kind of network structure comprising 3 layers of forward direction type of input layer, hidden layer and output layer.The input space transforms to hidden layer space through Nonlinear Mapping, and hidden layer space transforms to output layer space through linear mapping again.Wherein, the nonlinear mapping function that hidden layer adopts is radial basis function, and it is a kind of non-negative nonlinear function to the decay of central point radial symmetry of local distribution.
These neuron models have n input, h hidden node, and m the RBF network exported, namely this network structure is n-h-m.Wherein, x=(x
1, x
2..., x
n)
t∈ R
nfor net input vector, Φ
i() is the activation function of i-th hidden layer node, W ∈ R
h × mfor exporting weight matrix.Σ in figure in output layer node represents that output layer neuron adopts linear activation function.
Because the form of Gaussian function is relatively simple, analyticity is good, is convenient to carry out theoretical analysis, and significantly can not increase the complexity of calculating when multivariate inputs, and slickness is relatively good, and the derivative of Any Order all can exist, and radial symmetry.So, select Gaussian function as radial basis function herein.Now, the output of hidden layer and node i can be expressed as:
In formula, x is that d ties up input vector; σ
ibe the normalized factor of i-th basis function, it characterizes the width of this base letter; c
ithe center of i-th basis function, identical with x dimension; || x-c
i|| be x-c
inorm, i.e. x and c
ibetween distance.Φ
ix () is at c
iplace has and only has a maximal value, and with || x-c
i|| increase, Φ
ix () decays to zero rapidly; M is node in hidden layer, and so the output of radial basis function neural network is:
By the above Forecasting Methodology to wind power and load, if hour to calculate, we can obtain wind power and the payload in each hour moment, so when carrying out risk assessment to the electrical network of consideration wind power and load fluctuation, 24 assessments will be carried out at one day, when risk assessment each time, the situation after to each line fault is all needed to analyze, in the face of so complicated, during the calculating of magnanimity, traditional methods of risk assessment is due to cannot the importance degree of Complete Characterization different elements, calculate more complicated, its error is larger, existing network is actual to cause assessment result accurately not reflect, extra manpower and materials loss can be caused in implementation process, meanwhile, after wind-powered electricity generation is connected to the grid, have undulatory property and uncertainty, the running status of system is in the state of a change always, and its security risk is also in change, and prior art does not have a good method to assess wind-powered electricity generation security risks.
Therefore the present invention also provides a kind of methods of risk assessment, concrete, referring to Fig. 2, comprises the steps:
According to the importance degree of element each in electrical network, set up respectively and characterize grid nodes, the pitch point importance factor of branch road importance degree, branch road importance factors.
Concrete, the most complication system in reality can describe by the form of network.Wherein, electric system is typical complicated nonlinear system, utilizes Complex Networks Theory, and in conjunction with electrical network characteristic, power system network can be reduced to by N number of node, and the oriented network of having the right of M bar limit composition, its interior joint is divided into generating, load and contact node three class.
In traditional Complex Networks Theory, node and branch road important attribute is in the network architecture characterized by setting up betweenness (betweenness) index, this index hypothesis trend is only transmitted by shortest path between two nodes, and this hypothesis does not obviously meet reality to power engineering.Electric betweenness (electricbetweenness) index definition that the present invention is based on the proposition of Kirchhoff law is as follows.
The electric betweenness B of node n
en () is defined as:
In formula: G, L are respectively generator node set and load bus set; (i, j) is all " generating-load " nodes pair; w
ifor the weight of generator node i, get generator rated capacity or actually to exert oneself; w
jfor the weight of load bus j, get reality or peak load; B
e, ijn () is the electric betweenness produced on node n after adding Injection Current unit of unit between (i, j).
In formula: I
ij(m, n) is the electric current caused on branch road m-n add Injection Current unit of unit between (i, j) after; M is all nodes having branch road to be directly connected with n.
Same, the electric betweenness B of branch road l
el () is defined as:
In formula: I
ijl () is add unitary current unit between " generating-load " node is to (i, j) after, the electric current that branch road l causes.
Wherein,
represent respectively and characterize grid nodes, the pitch point importance factor of branch road importance degree, branch road importance factors.
Introduce described branch road importance factors, revise the branch road overload severity function of network system, introduce the described pitch point importance factor, revise the node low-voltage severity function of network system.
Concrete, the overload severity function of corresponding branch road j is:
In formula, P
jfor the active power of branch road j current transmission, P
limfor branch road overload risk max-thresholds, be set as the active power ultimate value that circuit is fully loaded herein.P
dfor the branch road overload risk threshold value of setting, generally get P
lim90%.
For the otherness of the running status order of severity of different elements in reasonable characterization system, introduce above-mentioned branch road importance factors as weight factor, the circuit overload severity of the define system overall situation is:
Wherein j represents jth bar branch road, B
e(l
j) representing branch road importance factors, j is positive integer.
Concrete, the node low-voltage severity function of corresponding node i is:
In formula, V
ifor the working voltage that node i is current, V
nfor node voltage ratings, V
limfor the low-voltage greateset risk threshold value of setting, be generally set as 90% of rated voltage.
For the otherness of the running status order of severity of different elements in reasonable characterization system, using the pitch point importance factor as weight factor, the low-voltage severity of the define system overall situation is:
Wherein i represents i-th node, B
e(l
j) representing the pitch point importance factor, i is positive integer.
The severity index characterizing operation of power networks state is obtained according to described branch road overload severity function, described node low-voltage severity function.
Concrete, undertaken comprehensively by node low-voltage severity function and overload severity function, can obtain the severity index of operation of power networks state after sign accident, expression formula is:
S
ev＝S(V)+S(P)
Introduce load economic factor, obtain the mistake load severity function characterizing load loss.
Concrete, in electric system, element fault often can cause the loss of load, and in safety evaluation, load loss mainly contains following three types:
Certain line fault system that directly causes out of service loses associated loadings node, and now load loss amount is this node load amount;
Low-voltage load sheding device action, after load bus busbar voltage is reduced to setting value, the load of excision setting;
System sectionalizing is after several isolated island, for keeping each isolated island power-balance, need add the mistake load that relevant control measure causes.
Because load attribute in electric system is different, for embodying the otherness of the different load loss order of severity, introduce load economic factor in mistake load evaluation index, definition load loss ratio is:
In formula, η is system loading loss ratio after accident, and L is load bus set, and L' is for losing load bus set, ε
ifor the economic factor of load bus i, p
ifor losing the load loss amount of load bus i, P
jfor the load of accident preload node j.
Load severity function is lost in definition:
In formula, η
limfor system loading loss maximum set threshold value, the present invention is taken as 20% of network load total amount.
Respectively process is weighted to described operation of power networks severity index, mistake load severity function, obtains the comprehensive severity function characterizing power grid accident.
Concrete, damage sequence severity not only comprises the impact of accident on system run all right, and the load loss of system has substantial connection simultaneously and after accident.The above-mentioned operation of power networks state severity index in conjunction with network structure and running status and load loss severity function are weighted process, obtain the comprehensive severity function after power grid accident, as shown by the equation:
S＝α×S
ev+β×S
load
＝α×[S(V)+S(P)]+β×S
load
Wherein α, β are that operation of power networks severity weight and electrical network lose load severity weight, and α, β value is respectively 0.3 and 0.7 in the present invention.
Further, also comprise: obtain integrated risk index according to described comprehensive severity function and contingency occurrence probability, and according to described index, risk assessment is carried out to network system.
Concrete, can find out that the probability that electric system has an accident meets Poisson (Poisson) distribution substantially from accident statistics, namely
In formula: E
ibe i-th systematic failures; P (E
i) be accident E
ithe probability occurred; λ
ifor the failure rate of i element in system.
Risk assessment considers the order of severity after accident possibility occurrence and accident.Comprehensive severity function after power grid accident obtained above is introduced in risk assessment index by the present invention, and after obtaining electric network fault, security of system integrated risk evaluation index is:
R
i＝P(E
i)·S
＝P(E
i)·{α×[S(V)+S(P)]+β×S
load}
Methods of risk assessment of the present invention, from methodology angle, electric betweenness is adopted to portray electrical network interior joint and branch road significance level in systems in which, overcome in traditional methods of risk assessment for the deficiency that the importance degree of different elements cannot characterize, press close to electrical network reality, make risk evaluation result more accurately reliable, thus reduce extra manpower and materials loss.In this appraisal procedure, introduce evaluation time window simultaneously, consider that the fluctuation of wind power output and load level is on the impact of system running state, establishes the electrical network dynamic secure estimation method considering blower fan access.
Concrete, the whole flow process of the present invention comprises the following steps:
(1), wind energy turbine set raw data and load raw data is gathered;
(2), analyze wind energy turbine set raw data, set up wind power seasonal effect in time series chaos phase space according to (1), and its chaotic characteristic is identified;
(3), the iteration result of (2)-(5) is used to carry out cluster to the phase point in phase space;
(4), according to (11)-(13) in every class phase point, screen the set of future position neighbor point, and limit the number of neighbor point according to (14);
(5), according to (6), improvement local Volterra sef-adapting filter model is set up to neighbor point;
(6) predicted data of wind energy turbine set, is obtained by the result of (6);
(7), analysis load raw data, set up RBF neural model according to (16) and (17), and obtain the predicted data of load;
(8), in power grid risk assessment time window select evaluation time point, and determine that the blower fan of this time point is exerted oneself and load level according to the short term predicted data of wind energy turbine set and load;
(9), to the electrical network that seclected time puts N-1 risk assessment is carried out, and selected assessment circuit;
(10), judge the network connectivty of method after selected line fault whether to produce isolated node; If without isolated node, jump to step (12); If generation isolated node, jump to step (11);
(11) load loss of isolated node, is calculated;
(12) network topology structure parameter, is set up; Above-mentioned steps (10)-(12) all belong to prior art, do not repeat them here;
(13), the importance factors of each element in computing method; Its circular describes see content of the present invention;
(14), Load flow calculation is carried out to current method; Existing mature technology is also belonged to for method Load flow calculation, no longer describes in detail;
(15), judged whether voltage out-of-limit, the whether action of low-voltage load sheding device, if low-voltage load sheding device action, calculated load is lost, and skips to step (14);
(16), according to the component importance factor of current system conditions and step (13), the probability level of system state and severity index are calculated, obtains the integrated risk index under system current failure; Its circular describes see content of the present invention;
(17), judge whether to travel through all circuits in electrical network, if do not complete traversal, skip to step (9); If complete traversal, find out the N-1 maximum risk value of electrical network on this time point;
(18), judge whether to carry out N-1 risk assessment to the electrical network of all time points, if do not completed, skip to step (8);
(19), the N-1 fault maximum risk value variation tendency chart of electrical network in evaluation time window is exported;
(20), terminate.
The present invention proposes and a kind ofly consider that the electric network security methods of risk assessment of evaluation time window carries out dynamic secure estimation to the electrical network accessed containing blower fan.The method can be exerted oneself and power grid risk level in load level situation by the different blower fan of Efficient Characterization, and can provide the electrical network N-1 fault maximum risk value variation tendency in whole evaluation time window; The risk profile appraisal procedure of a kind of wind power integration electrical network of the present invention, the risk of each time point is carried out comprehensively, utilize network system risk to walk power curve and for system cloud gray model personnel prediction and system running state can be grasped, can be good at predicting wind power, thus effectively prevention to be fluctuated the power system accident caused due to wind-powered electricity generation, avoids causing large-area power-cuts.
By reference to the accompanying drawings the specific embodiment of the present invention is described in detail above, but the present invention is not restricted to above-mentioned embodiment, in the spirit and scope situation of claim not departing from the application, those skilled in the art can make various amendment or remodeling.
Claims (5)
1. a risk profile appraisal procedure for wind power integration electrical network, is characterized in that, comprise the following steps:
Gather wind energy turbine set raw data, obtain initial wind power time series;
Phase space reconfiguration is carried out to described initial wind power time series, obtains phase point vector;
The calculating of K mean cluster is carried out to described phase point vector;
Introduce weight vectors and improve the distance of neighbor point and the criterion of trend, set up comprehensive criterion, obtain neighbor point set;
Using the training set of described neighbor point set as volterra sef-adapting filter, obtain wind farm power prediction data.
2. the risk profile appraisal procedure of a kind of wind power integration electrical network according to claim 1, is characterized in that, described introducing weight vectors improves the distance of neighbor point and the criterion of trend, sets up comprehensive criterion, obtains neighbor point set and comprise:
Define a kind of compute mode:
Wherein A, B are the vector of m dimension, a
i, b
ibe respectively the i-th dimension component of vectorial A, B;
Distance then between current predictive point X (p) and phase point X (i) is:
In formula, X (p) represents p future position, X (i) represents i-th phase point, α is weight vectors, and for the vectorial α of m dimension, α (1)≤α (2)≤... ≤ α (m), considers that interval time between coordinate components is for being τ, gets
Wherein, d (p, i) is less, and current predictive point X (p) is nearer with the distance of phase point X (i);
The difference value vector of definition multistep backtracking is:
E(p,q)＝X(p)-X(p-q)，
E(i,q)＝X(i)-X(i-q)，
Wherein, q represents backtracking step-length, and the q that X (p) is X (p-q) walks evolution phase point, and the q that X (i) is X (i-q) walks evolution phase point;
Angular separation between the future position of multistep backtracking and phase point is:
Be weighted above formula, the development trend criterion that can obtain future position and phase point is:
Cos θ (p, i) be by vector between folder cosine of an angle develop, wherein, β is weight vectors, and β (0)≤β (1)≤... ≤ β (q), wherein
Cos θ (p, i) is less, represents that current predictive point X (p) is more close with the development trend of phase point X (i);
Similarity between future position X (p) and phase point X (i) is:
η(p,i)＝γ
1d(p,i)+γ
2cosθ(p,i)，
Wherein γ
1, γ
2be respectively the weighted value of range index and evolution trend index, and γ
1+ γ
2=1.
3. the risk profile appraisal procedure of a kind of wind power integration electrical network according to claim 1 and 2, is characterized in that, also comprise:
Wind energy turbine set load data, obtains load prediction data according to RBF neural model;
Wind power and predicted load is obtained according to described wind farm power prediction data and load prediction data;
Selected multiple evaluation time point, carries out risk assessment according to the wind power of each described evaluation time section and predicted load to electrical network, obtains multiple risk evaluation result;
Power grid risk prediction curve is obtained according to described multiple risk evaluation result.
4. the risk profile appraisal procedure of a kind of wind power integration electrical network according to claim 3, is characterized in that, the described wind power according to each described evaluation time section and predicted load carry out risk assessment to electrical network and comprise:
According to the importance degree of element each in electrical network, set up respectively and characterize grid nodes, the pitch point importance factor of branch road importance degree, branch road importance factors;
Introduce described branch road importance factors, revise the branch road overload severity function of network system, introduce the described pitch point importance factor, revise the node low-voltage severity function of network system;
The severity index characterizing operation of power networks state is obtained according to described branch road overload severity function, described node low-voltage severity function;
Introduce load economic factor, obtain the mistake load severity function characterizing load loss;
Respectively process is weighted to described operation of power networks severity index, mistake load severity function, obtains the comprehensive severity function characterizing power grid accident.
5. the risk profile appraisal procedure of a kind of wind power integration electrical network according to claim 4, it is characterized in that, also comprise: obtain integrated risk index according to described comprehensive severity function and contingency occurrence probability, and according to described index, risk assessment is carried out to network system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510824706.4A CN105354643A (en) | 2015-11-24 | 2015-11-24 | Risk prediction evaluation method for wind power grid integration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510824706.4A CN105354643A (en) | 2015-11-24 | 2015-11-24 | Risk prediction evaluation method for wind power grid integration |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105354643A true CN105354643A (en) | 2016-02-24 |
Family
ID=55330610
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510824706.4A Pending CN105354643A (en) | 2015-11-24 | 2015-11-24 | Risk prediction evaluation method for wind power grid integration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105354643A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913160A (en) * | 2016-05-09 | 2016-08-31 | 国网四川省电力公司经济技术研究院 | Calculation method capable of employing transmission capability based on large-scale wind power integration |
| CN106347578A (en) * | 2016-10-27 | 2017-01-25 | 江苏科技大学 | Method for monitoring marine propulsion shafting bearing vibration failure |
| CN106950421A (en) * | 2017-01-03 | 2017-07-14 | 内蒙古电力(集团)有限责任公司 | A kind of active power for wind power extremum extracting method and device |
| CN107317334A (en) * | 2017-08-31 | 2017-11-03 | 华北电力大学（保定） | A kind of power system rack reconstructing method and device |
| CN108898262A (en) * | 2018-07-26 | 2018-11-27 | 国网山东省电力公司莱芜供电公司 | A kind of wind power prediction method considering wind-powered electricity generation uncertainty risk |
| CN109146751A (en) * | 2018-08-29 | 2019-01-04 | 昆明理工大学 | A method of city crime risk assessment is carried out based on space-time class statistic |
| CN109255730A (en) * | 2018-11-13 | 2019-01-22 | 中石化石油工程技术服务有限公司 | Load forecasting method when a kind of natural gas |
| CN109344995A (en) * | 2018-08-29 | 2019-02-15 | 广东工业大学 | A kind of chaos time sequence multistep forecasting method based on density peaks cluster |
| CN112508276A (en) * | 2020-12-07 | 2021-03-16 | 国网河南省电力公司洛阳供电公司 | Power grid rapid diagnosis and optimization system and optimization method |
| CN113705874A (en) * | 2021-08-19 | 2021-11-26 | 国家电网有限公司 | New energy power grid evolution prediction method and device, computer equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102685766A (en) * | 2012-05-13 | 2012-09-19 | 西华大学 | Wireless network flow prediction method based on local minimax probability machine |
| CN102842105A (en) * | 2012-07-09 | 2012-12-26 | 中国电力科学研究院 | Online transient state stability risk evaluating method for metering wind power uncertainty |
| CN103248051A (en) * | 2013-05-25 | 2013-08-14 | 南京南瑞集团公司 | Method for evaluating grid operation safety risk caused by wind farm power fluctuation |
| CN103246806A (en) * | 2013-04-25 | 2013-08-14 | 浙江大学 | Operation risk evaluation method comprising wind- power plant electric system |
-
2015
- 2015-11-24 CN CN201510824706.4A patent/CN105354643A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102685766A (en) * | 2012-05-13 | 2012-09-19 | 西华大学 | Wireless network flow prediction method based on local minimax probability machine |
| CN102842105A (en) * | 2012-07-09 | 2012-12-26 | 中国电力科学研究院 | Online transient state stability risk evaluating method for metering wind power uncertainty |
| CN103246806A (en) * | 2013-04-25 | 2013-08-14 | 浙江大学 | Operation risk evaluation method comprising wind- power plant electric system |
| CN103248051A (en) * | 2013-05-25 | 2013-08-14 | 南京南瑞集团公司 | Method for evaluating grid operation safety risk caused by wind farm power fluctuation |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105913160A (en) * | 2016-05-09 | 2016-08-31 | 国网四川省电力公司经济技术研究院 | Calculation method capable of employing transmission capability based on large-scale wind power integration |
| CN105913160B (en) * | 2016-05-09 | 2019-12-03 | 国网四川省电力公司经济技术研究院 | A kind of calculation method using transmittability based on large-scale wind power integration |
| CN106347578A (en) * | 2016-10-27 | 2017-01-25 | 江苏科技大学 | Method for monitoring marine propulsion shafting bearing vibration failure |
| CN106950421B (en) * | 2017-01-03 | 2019-04-12 | 内蒙古电力(集团)有限责任公司 | A kind of active power for wind power extremum extracting method and system |
| CN106950421A (en) * | 2017-01-03 | 2017-07-14 | 内蒙古电力(集团)有限责任公司 | A kind of active power for wind power extremum extracting method and device |
| CN107317334A (en) * | 2017-08-31 | 2017-11-03 | 华北电力大学（保定） | A kind of power system rack reconstructing method and device |
| CN108898262A (en) * | 2018-07-26 | 2018-11-27 | 国网山东省电力公司莱芜供电公司 | A kind of wind power prediction method considering wind-powered electricity generation uncertainty risk |
| CN108898262B (en) * | 2018-07-26 | 2021-08-06 | 国网山东省电力公司莱芜供电公司 | Wind power prediction method considering wind power uncertainty risk |
| CN109344995A (en) * | 2018-08-29 | 2019-02-15 | 广东工业大学 | A kind of chaos time sequence multistep forecasting method based on density peaks cluster |
| CN109146751A (en) * | 2018-08-29 | 2019-01-04 | 昆明理工大学 | A method of city crime risk assessment is carried out based on space-time class statistic |
| CN109344995B (en) * | 2018-08-29 | 2021-06-18 | 广东工业大学 | Multi-step prediction method of chaotic time sequence based on density peak clustering |
| CN109146751B (en) * | 2018-08-29 | 2021-10-22 | 昆明理工大学 | Method for urban crime risk assessment based on space-time clustering statistics |
| CN109255730A (en) * | 2018-11-13 | 2019-01-22 | 中石化石油工程技术服务有限公司 | Load forecasting method when a kind of natural gas |
| CN112508276A (en) * | 2020-12-07 | 2021-03-16 | 国网河南省电力公司洛阳供电公司 | Power grid rapid diagnosis and optimization system and optimization method |
| CN113705874A (en) * | 2021-08-19 | 2021-11-26 | 国家电网有限公司 | New energy power grid evolution prediction method and device, computer equipment and storage medium |
| CN113705874B (en) * | 2021-08-19 | 2024-03-15 | 国家电网有限公司 | New energy power grid evolution prediction method, device, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105354643A (en) | Risk prediction evaluation method for wind power grid integration | |
| Azman et al. | A unified online deep learning prediction model for small signal and transient stability | |
| Xu et al. | Assessing short-term voltage stability of electric power systems by a hierarchical intelligent system | |
| Bolognani et al. | Fast power system analysis via implicit linearization of the power flow manifold | |
| Zhao et al. | Resilient unit commitment for day-ahead market considering probabilistic impacts of hurricanes | |
| CN111628501B (en) | AC/DC large power grid transient voltage stability assessment method and system | |
| CN109993665B (en) | Online safety and stability assessment method, device and system for power system | |
| Wang et al. | Operating state reconstruction in cyber physical smart grid for automatic attack filtering | |
| CN111191832A (en) | Typhoon disaster power distribution network tower fault prediction method and system | |
| Wang et al. | Data-driven prediction method for characteristics of voltage sag based on fuzzy time series | |
| Zhu et al. | Integrated data-driven power system transient stability monitoring and enhancement | |
| Duan et al. | Security risk assessment using fast probabilistic power flow considering static power-frequency characteristics of power systems | |
| Gong et al. | Transient stability assessment of electric power system based on voltage phasor and cnn-lstm | |
| Rayudu | A knowledge-based architecture for distributed fault analysis in power networks | |
| CN117335426A (en) | Key feature identification method suitable for short-term voltage stability evaluation | |
| Karapidakis | Machine learning for frequency estimation of power systems | |
| Sun et al. | Reliability prediction of distribution network based on PCA-GA-BP neural network | |
| JP6818658B2 (en) | Power system monitoring system, power system monitoring method, and program | |
| Yang et al. | Research on CNN-LSTM DC power system fault diagnosis and differential protection strategy based on reinforcement learning | |
| Ye et al. | Switching Convolution of Node Graph and Line Graph-driven Method for Fast Static Security Analysis | |
| Alquthami et al. | Forecasting Transmission System Outages Using Artificial Neural Networks | |
| Liu et al. | An online power system transient stability assessment method based on graph neural network and central moment discrepancy | |
| Xu et al. | Evaluation of Distribution Line Fault Probability and Line Reinforcement Method in Typhoon Events | |
| Putranto et al. | Series PMU data-based state estimation technique for WAMS application | |
| Jiang et al. | Research on power grid fault diagnosis technology based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160224 |