CN109472428A

CN109472428A - A risk assessment method for distribution network operation based on loss expectation method

Info

Publication number: CN109472428A
Application number: CN201810070384.2A
Authority: CN
Inventors: 吴宇红; 纪涛; 周健; 赖旬阳; 李洋; 杨强; 薄耀龙; 郑军
Original assignee: Zhejiang Deqing County Power Supply Co Ltd; Zhejiang University ZJU; Huzhou Power Supply Co of State Grid Zhejiang Electric Power Co Ltd; State Grid Corp of China SGCC
Current assignee: Zhejiang Deqing County Power Supply Co Ltd; Zhejiang University ZJU; Huzhou Power Supply Co of State Grid Zhejiang Electric Power Co Ltd; State Grid Corp of China SGCC
Priority date: 2018-01-24
Filing date: 2018-01-24
Publication date: 2019-03-15

Abstract

The invention relates to a distribution network operation risk assessment method based on the loss expectation method. The main problem is that the prior art is immature for distribution network operation risk assessment. The invention fully considers various influencing factors in the operation process of the distribution network. Based on the loss expectation method, the product of probabilistic item score, risk item score and load important factor is proposed to represent the specific risk of distribution network. The invention reduces the subjectivity of weight determination and increases the objectivity; meanwhile, the difference and characteristics of different distribution network topologies for distribution network risks are considered, so that risk assessment is more consistent with reality. The invention takes the modern distribution network model as the consideration object, lists the risk indicators, fully considers the equipment level, operation and maintenance level, intelligence level, anti-external force damage level and other indicators of the distribution network, and has the advantages of small calculation amount, high accuracy and low cost. The weak links of the distribution network have a deep understanding and are suitable for increasingly large distribution network systems.

Description

A kind of power distribution network operation risk assessment method based on loss expectation method

Technical field

The present invention relates to electrical engineering and power distribution network risk assessment field, more particularly, to a kind of based on loss expectation method Power distribution network operation risk assessment method.

Background technique

Power grid is the important composition portion of national infrastructure as one of engineering project extremely complex in human history Point.Guarantee continue, efficient power supply be electric power enterprise first have to consider target, and society work normally life must Want condition.Currently, China has been realized in large-scale networking, northeast and North China, North China and Central China, Central China and East China, Central China Interconnection is all had been realized in northwest, Central China and south electric network.Especially at the end of the 20th century, information is rapidly sent out with Internet technology Under the prospect of exhibition, Chinese electricity power engineering has been achieved for many good achievements.But, it has to it is mentioned that, in China, The thought of " retransmitting light defeated ineffective " still has, and this thought causes now due to occupying leading position in early days As the scale and structure of electric system are increasingly complicated and huge, the uncertain influence of electric system is also more and more.According to National urban customer power supply reliability index situations in 2010 of national Electricity Monitoring Commission's management of electric power dependability center publication, remove system The conditions of rationing the power supply such as off-capacity of uniting, the power failure overwhelming majority that user is subjected to account for about total power failure the reason is that as caused by distribution link 96% or so of event, and among these, the more accounting examples of mesolow distribution network systems are more up to 90%.These fearful numbers are obvious Reflect, structural instability in China's distribution network system, there is an urgent need to be transformed.Therefore it needs to comment power distribution network progress risk Estimate, on the one hand in order to find the weak link in current power distribution network, on the other hand can be used to instruct the following intelligent distribution network Construction, risk assessment have great importance.

The difficult point of power distribution network risk assessment is: determine value-at-risk of which type of standard as power distribution network, i.e., it is how right The risk of power distribution network is quantified；More perfect, comprehensive, reasonable, objective appraisal system and model are how established, is answered In the risk assessment for using Modern power distribution net；How objective reasonably determine is related to the index weights of risk projects, has obtained More accurate reasonable assessment result.Present document power distribution network risk assessment has been made very more explorations, proposes The method of many power distribution network risk assessment: the model risk method for tracing based on well-being, the shadow of comprehensive weather conditions It rings, the sequence and assessment of component importance has been carried out to power distribution network key equipment；Foundation is able to reflect electric system real time execution The element time-varying reliability model of condition, the operation risk index system for constructing detailed characterizations reliability in time level；

The deficiency of research, which is 1. to study, to be concentrated mainly in economic risk, could not be carried out to fault outage risk for research； 2. the difference between fault outage and pre-arranged power failure could not be distinguished strictly；3. too emphasizing the real-time of power distribution network risk, i.e., Consider power networks risk in a short time, rather than is analyzed in conjunction with historical data.

Application No. is 201510026092.5, a kind of entitled " power distribution network operation risk assessment based on average expected volume A kind of power distribution network operation risk assessment method based on average expected volume is disclosed in the patent specification of method ", the invention is not The difference between fault outage and pre-arranged power failure can be strictly distinguished, Evaluated effect is not comprehensive.

Summary of the invention

The present invention mainly solve prior art research be concentrated mainly in economic risk, cannot to fault outage risk into The hand-manipulating of needle to research, cannot strictly distinguish fault outage and pre-arranged have a power failure between difference and too emphasize power distribution network risk The problem of real-time considers power networks risk in a short time, rather than combination historical data is analyzed, provides one Power distribution network operation risk assessment method of the kind based on loss expectation method.

Above-mentioned technical problem of the invention is mainly to be addressed by following technical proposals: one kind is based on loss expectation The power distribution network operation risk assessment method of method, comprising the following steps:

S1. it determines to the contributive project indicator of power distribution network risk；

S2. the electrical circuitry equipment of power distribution network, network structure, operation and maintenance record, failure logging, responsible consumer etc. are collected respectively Basic data, determine probability project indicator value and consequent project indicator value；

S3. risk scale is set with analytic hierarchy process (AHP), combines consistency judge index model after obtaining judgment matrix, determined each The weight coefficient of project；

S4. Delphi method method is used, each project evaluation is evaluated, determines that standards of grading, each index are commented Minute mark is quasi- are as follows: minimum 0 point, highest 100 is divided, and for certain project indicators, index value more balloon score is lower；For other projects Index, index value is lower, and score is higher；And some projects are that there are an interval values, i.e., the index value in this section is corresponding Score be equal；

S5. according to minimal path block algorithm in power distribution network, power distribution network is divided into multiple independent pieces, is to include in each piece The user of multiple and different load levels, can be obtained by the particle calculation formula in physical model, the load in power distribution network it is important because Element is equivalent to the particle weight in network, can obtain load key factor, it may be assumed that

Wherein, S_iFor i-th of piecemeal,Z indicates the load important factor in power distribution network；

S6. it can be calculated power distribution network operation overall risk R, it may be assumed that

Wherein, S_proIndicate probability project indicator score, S_conIndicate consequent project indicator score, Z is negative lotus important factor.

The present invention has fully considered each influence factor in power distribution network operational process, in conjunction with the development of up-to-date technology to matching Power grid risk project is further optimized, and proposes based on the probability item score of loss expectation method, risky projects point Counting with load key factor three product indicates the specific risk of power distribution network.Specific power distribution network is primarily based on for each project Determine its index value, then application level analytic approach determines the weight coefficient W of each project_i, then use Delphi method It scores each project, determines standards of grading, finally according to minimal path distribution network load split plot design, determine and be based on power distribution network The load important factor of topology.Present invention reduces the subjectivities that weight determines, increase objectivity；Difference is considered simultaneously Difference and feature of the distribution net topology for power distribution network risk so that risk assessment and practical more consistent.

As a preferred embodiment, in step S1 the project indicator determination process are as follows:

S11. fish-bone chart is applied, being with " people, method, material, machine, ring " greatly will be because carrying out risk identification to power distribution network；

S12. the particularity and common point for fully considering different power distribution networks establish reasonable and applicable venture influence element；

S13. gone out by above-mentioned venture influence element comprehensive analysis to the contributive project indicator of power distribution network risk.

As a preferred embodiment, each probability project indicator value determined in step S2 are as follows:

As a preferred embodiment, each consequent project indicator value determined in step S2 are as follows:

As a preferred embodiment, step S3 the following steps are included:

S31. all characteristic items involved by the problem can clearly be reacted by establishing one, form a recursive hierarchy structure, i.e. A It include B1, B2, B3 ... in A for top layer project, Bn total n subitem contains C1, C2, C3 ... in B1, the total m subitem of Cm, Contain D1, D2, D2 ... in B2, the total t subitem of Dt, and so on, each top layer project can be divided into multiple sublayer items Mesh；

S32. corresponding proportion quotiety and definition, i.e., the relationship of mutual importance and quantitative value between element, setting mark are provided Degree is 1,2,3,4,5,6,7,8,9,10, wherein scale 1 indicates that a is important as b, and scale 2 indicates that a ratio b is slightly important, mark Degree 4 indicates that a and b is generally important, and scale 6 indicates that a ratio b is more important, and scale 8 indicates that a ratio b is extremely important, and scale 10 indicates a ratio B is incomparably important, and scale 3,5,7,9 indicates the value for judging more to obscure between scale, can select above-mentioned scale accordingly；

S33. by comparing the relative importance between element in each layer, a certain subitem to lower layer in relation to this in upper layer is constructed Judgment matrix, it may be assumed that

Wherein, P_ijIndicate that i-th of index to the importance quantitative values of j-th of index, and defines,

S34. judgment matrix is handled, lower layer's subitem is obtained and the relative importance on upper layer is arranged.In processing array, generally It is to calculate the corresponding characteristic root of a matrix, is normalized to obtain weight coefficient later.But due to judgment matrix inherently one A subjective model has comparable error range, and factor weight in the level that provides of application level analytic approach Coefficient is substantially the qualitative relationships of a priority.Therefore can be used ask approximate maximum eigenvalue and corresponding feature to Amount handles judgment matrix to obtain approximate solution using root method here, obtains lower layer's subitem to the relatively important journey on upper layer Degree arrangement, it may be assumed that the product T of each row element in judgment matrix P is first calculated,

Then the Nth power root of T is calculated, what N was indicated is the order of judgment matrix,

Finally vector W=[W1, W2 ..., Wn] T is normalized,

Vector W0=[W10, W20 ..., the Wn0] T then obtained is required feature vector, and each element is weight coefficient；

S35. the maximum eigenvalue of judgment matrix is calculated, it may be assumed that the corresponding feature vector PW of maximum eigenvalue is first calculated,

PW=P × W₀,

Maximum eigenvalue is obtained later:

S36. the consistency of judgment matrix is verified, it may be assumed that coincident indicator CI and consistency ratio CR are respectively as follows:

Wherein, RI is the Aver-age Random Consistency Index of multistage judgment matrix, is constant relevant to N, establishes judgment matrix When, when the element on same level is more, may occur self-contradictory situation in the judgment process, in general, judgement Order of matrix number is higher, and the difficulty that judgment matrix remains exactly the same is bigger, wherein share 10 ranks, respectively 1 rank, 2 ranks, 3 ranks, 4 ranks, 5 ranks, 6 ranks, 7 ranks, 8 ranks, 9 ranks, 10 ranks, corresponding RI value is respectively 0,0,0.58,0.9,1.12,1.24, 1.32,1.41,1.45,1.49, as CR < 0.10, just thinks that judgment matrix has satisfactory consistency, otherwise need to adjust judgement Matrix makes it meet CR < 0.10, until reaching with satisfied consistency.After obtaining satisfied consistency, analytic hierarchy process (AHP) Also analysis finishes, and more satisfied weight coefficient can be obtained.

As a preferred embodiment, in step S5 calculated load key factor process the following steps are included:

S51. determining autonomous block is split to power distribution network, similar load is placed in same piece；

S52. the power distribution network node set of all branches (line and transformer branch etc.) is set as S_b0, S_bFor S_b0A subset, S_nbFor with S_bRelevant node collection；

S53. the nodal analysis method of power distribution network is traversed, finds the shortest path of each load and the block at place；

S54. the particle calculation formula in analogy physical model, the load key factor in power distribution network are equivalent to the particle in network Weight, it may be assumed that

Wherein, S_iFor i-th of piecemeal,Z indicates the load important factor in power distribution network.

Therefore, the invention has the advantages that the present invention with Modern power distribution pessimistic concurrency control be consider object, to risk indicator carry out sieve Column, fully take into account the indexs such as equipment level, O&M level, intelligent level, the external force resistance destruction level of power distribution network, have meter Calculation amount is smaller, understands the weak link of power distribution network level depth, suitable for increasingly huger distribution network system.

Detailed description of the invention

Fig. 1 is a kind of processing flow schematic diagram of the invention；

Fig. 2 is minimal path power distribution network partitioning algorithm schematic diagram of the invention.

Specific embodiment

Below with reference to the embodiments and with reference to the accompanying drawing the technical solutions of the present invention will be further described.

Embodiment:

A kind of power distribution network operation risk assessment method based on loss expectation method of the present embodiment, as shown in Figure 1, including following step It is rapid:

S2. electrical circuitry equipment, network structure, operation and maintenance record, failure logging, the responsible consumer etc. of power distribution network are collected respectively The basic data of aspect determines probability project indicator value and consequent project indicator value；

Probability project indicator value:

Consequent project indicator value:

S3. risk scale is set with analytic hierarchy process (AHP), combines consistency judge index model after obtaining judgment matrix, really Determine the weight coefficient of projects；

S32. corresponding proportion quotiety and definition, i.e., the relationship of mutual importance and quantitative value between element, setting mark are provided Degree is 1,2,3,4,5,6,7,8,9,10, wherein scale 1 indicates that a is important as b, and scale 2 indicates that a ratio b is slightly important, mark Degree 4 indicates that a and b is generally important, and scale 6 indicates that a ratio b is more important, and scale 8 indicates that a ratio b is extremely important, and scale 10 indicates a ratio B is incomparably important, and scale 3,5,7,9 indicates the value for judging more to obscure between scale, can select above-mentioned scale accordingly, such as Shown in following table:

Scale	The meaning of representative
		1	A is important as b
2	A ratio b is slightly important
		4	A ratio b is general important
6	A ratio b is more important
		8	A ratio b is extremely important
10	A ratio b is incomparable important
		3,5,7,9	Value between scale, if it is determined that more fuzzy can choose these values

Finally vector W=[W1, W2 ..., Wn] T is normalized,

PW=P × W₀,

Maximum eigenvalue is obtained later:

Wherein, RI is the Aver-age Random Consistency Index of multistage judgment matrix, is constant relevant to N, establishes judgment matrix When, when the element on same level is more, may occur self-contradictory situation in the judgment process, in general, judgement Order of matrix number is higher, and the difficulty that judgment matrix remains exactly the same is bigger, wherein share 10 ranks, respectively 1 rank, 2 ranks, 3 ranks, 4 ranks, 5 ranks, 6 ranks, 7 ranks, 8 ranks, 9 ranks, 10 ranks, corresponding RI value is respectively 0,0,0.58,0.9,1.12,1.24, 1.32,1.41,1.45,1.49, following table is the relationship table of RI value and order:

Order	1	2	3	4	5	6	7	8	9	10
											RI	0	0	0.58	0.9	1.12	1.24	1.32	1.41	1.45	1.49

As CR < 0.10, just think judgment matrix have satisfactory consistency, otherwise need to adjust judgment matrix, make its meet CR < 0.10, until reaching with satisfied consistency.After obtaining satisfied consistency, analytic hierarchy process (AHP) is also analyzed and is finished The weight coefficient being more satisfied with.

S4. Delphi method method is used, each project evaluation is evaluated, determines standards of grading, each index Standards of grading are as follows: minimum 0 point, highest 100 is divided, and for certain project indicators, index value more balloon score is lower；For other The project indicator, index value is lower, and score is higher；And some projects are that there are an interval values, i.e. the index value in this section Corresponding score is equal, and specific standards of grading and index weights are as shown in the table:

S51. as shown in Fig. 2, being split determining autonomous block to power distribution network, similar load is placed in same piece, and it is right After same piece, any one element fault, the switch and coverage acted in system is identical, it is possible to Regarded as a separate component；

S52. the power distribution network node set of all branches (line and transformer branch etc.) is set as S_b0, S_bFor S_b0A subset, S_nbFor with S_bRelevant node collection, in power distribution network, overstepping one's bounds branch line endpoint node is known as feed connection node.And if certain feed connection node (containing sub- feed connection node) is not connected with other any nodes, then the node is known as feeder terminal node.If S_n0For a distribution Net node set, S_n1For S_n0A proper subclass, then definition belong to S_n1And with The connected node of middle node is known as S_n1Inner boundary node.In power distribution network, if all branches (line and transformer branch etc.) Collection is combined into S_b0, S_bFor S_b0A subset, S_nbFor with S_bRelevant node collection.In addition to S_nbInner boundary node side outside, if S_b In any switchgear (breaker, fuse, block switch, disconnecting switch etc.) is not configured, then claim S_bFor member block.In element There is no switchgear in block, therefore in block after any element fault, the shadow of the sequence of movement and failure of breaker and disconnecting switch It is consistent to ring range；

Wherein, S_proIndicate probability project indicator score, S_conIndicate consequent project indicator score, Z is negative lotus important factor； The result of risk assessment is analyzed, rather than is solely focused on the size of overall risk, it is available more about power distribution network Information, to provide emphasis for later Distribution system design, and find the weak place of current power distribution network, carry out specific aim at Reason.

For probability item score, this paper defines probability item score and maximum value is 100, minimum value 0.When Score and be greater than 100 when, according to 100 metering.It defines shown in specific risk following table:

It is especially big	It is larger	Generally	It is smaller	Very little
					>80	[50-80]	[20,50]	[20,20]	<20

For consequent item score, this paper defines consequent item score and maximum value is 100, minimum value 0.Work as score When with being greater than 100, according to 100 meterings.It defines shown in specific risk following table:

For last overall risk, defining overall risk maximum value is 10000, minimum value 0.When total risk value is greater than 10000, It is calculated according to 10000.Shown in the specific scale following table for defining overall risk:

It is especially big	It is great	It is larger	Generally	Slightly
					[4900,10000]	[1600,4900]	[400,1600]	[100,400]	<100

Specific embodiment described herein is only an example for the spirit of the invention.The technical field of the invention Technical staff can make various modifications or additions to the described embodiments or be substituted in a similar manner, but Without departing from the spirit of the invention or going beyond the scope defined by the appended claims.

Claims

1. a distribution network operation risk assessment method based on loss expectation method is characterized in that: comprise the following steps:

S1. Determine the project indicators that contribute to the risk of the distribution network;

S2. Collect basic data of power equipment, network structure, operation and maintenance records, fault records, important users, etc. of the distribution network respectively, and determine the index values of probabilistic items and consequential items;

S3. Use the analytic hierarchy process to set the risk scale, obtain the judgment matrix and combine the consistency judgment index model to determine the weight coefficient of each item;

S4. Use the Delphi method to evaluate the evaluation of each item and determine the scoring standard;

S5. According to the minimum path block algorithm in the distribution network, the distribution network is divided into multiple independent blocks, each block contains multiple users with different load levels, and the particle calculation formula in the physical model can be used to calculate Therefore, the important factor of the load in the distribution network is equivalent to the weight of the particle in the network, and the important factor of the load can be obtained, namely:

Among them, S _i is the ith block, Z represents the load importance factor in the distribution network;

S6. Calculate the total risk R of distribution network operation, namely:

Among them, S _pro represents the probabilistic item index score, S _con represents the consequential item index score, and Z is the load importance factor.

2. a kind of distribution network operation risk assessment method based on loss expectation method according to claim 1 is characterized in that the determination process of project index in step S1 is:

S11. Use the fishbone diagram method to identify the risk of the distribution network with "people, methods, materials, machines, and rings" as the main factors;

S12. Fully consider the particularity and commonality of different distribution networks, and establish reasonable and applicable risk factors;

S13. Comprehensively analyze the project indicators that contribute to the risk of the distribution network through the above risk influencing factors.

3. a kind of distribution network operation risk assessment method based on loss expectation method according to claim 1 is characterized in that each probabilistic item index value determined in step S2 is:

4. a kind of distribution network operation risk assessment method based on loss expectation method according to claim 1 is characterized in that each consequence item index value determined in step S2 is:

5. A kind of distribution network operation risk assessment method based on loss expectation method according to claim 1 is characterized in that step S3 comprises the following steps:

S31. Establish a feature item that can clearly reflect all the feature items involved in the problem and form a hierarchical structure;

S32. Give the corresponding scale and definition, that is, the relationship between the mutual importance of elements and the quantitative value;

S33. By comparing the relative importance of elements in each layer, construct a judgment matrix for a certain item in the upper layer to the sub-item of the sub-item in the lower layer, namely:

Among them, P _ij represents the quantitative value of the importance of the ith index to the jth index, and is defined as

S34. Use the square root method to process the judgment matrix, and obtain the relative importance arrangement of the lower-level sub-items to the upper-level items, that is: first calculate the product T of each row element in the judgment matrix P,

Then calculate the N-th root of T, where N represents the order of the judgment matrix,

Finally, normalize the vector W=[W1,W2,...,Wn]T,

Then the obtained vector W0=[W10, W20, ..., Wn0]T is the required eigenvector, and each element is the weight coefficient;

S35. Calculate the maximum eigenvalue of the judgment matrix, that is: first calculate the eigenvector PW corresponding to the maximum eigenvalue,

PW=P×W ₀ ,

Then get the largest eigenvalue:

S36. Check the consistency of the judgment matrix, that is, the consistency index CI and the consistency ratio CR are respectively:

where RI is the average random consistency index of the multi-order judgment matrix, and is a constant related to N.

6. a kind of distribution network operation risk assessment method based on loss expectation method according to claim 1 is characterized in that the process of calculating load important factor in step S5 comprises the following steps:

S51. Divide the distribution network to determine independent blocks, and place similar loads in the same block;

S52. Let the set of distribution network nodes of all branches (lines and transformer branches, etc.) be S _b0 , S _b be a subset of S _b0 , and _Snb be the node set related to S _b ;

S53. Traverse the node model of the distribution network to find the shortest path and block of each load;

S54. By analogy with the particle calculation formula in the physical model, the important factor of the load in the distribution network is equivalent to the weight of the particle in the network, namely:

Among them, S _i is the ith block, Z represents the load importance factor in the distribution network.