CN104715318A - Multi-dimensional operational risk evaluating method for communication network - Google Patents

Abstract

The invention discloses a multi-dimensional operational risk evaluating method for a communication network and relates to power grid operation risk evaluating methods. According to an existing electric power communication system monitoring system, the change of the network quality can not be represented truly, the guarantee capability of a system can not be evaluated comprehensively, and potential risks of the system can not be managed or controlled accurately. The multi-dimensional operational risk evaluating method for the communication network comprises the following steps that a multi-dimensional operational risk evaluating and early warning model for the communication network is established, the weights of evaluation indexes are determined, a grading rule for the evaluation indexes is determined, and comprehensive grading and managing are achieved through an electric power communication system. According to the multi-dimensional operational risk evaluating method for the communication network, a macroscopic result is reflected by evaluating the difference on the whole, the specific problem is accurately positioned and the future development trend and potential risks are predicted by means of multi-dimensional item evaluation, existing maintenance source investment efficiency and benefits can be evaluated in a quantitative mode, the method has great significance in guiding the total number of continuously-invested resources and the structural configuration, and an important basis is provided for daily management of the system.

Description

Towards the various dimensions operation risk evaluation method of communication network

Technical field

The present invention relates to operation of power networks risk evaluating method, especially towards the various dimensions operation risk evaluation method of communication network.

Background technology

Power communication system is the important foundation guaranteed power grid security, realize economic load dispatching, along with the development of system in network size, running quality, class of business etc., the difficulty of powerline network operation maintenance and risk management and control strengthens further, how effectively to tackle Outer risks change, promote integrated management level, become Utilities Electric Co. and pay close attention to present stage and the key issue explored.Meanwhile, numerous scholar also carries out continuing research to Study of Risk Evaluation Analysis for Power System method in recent years, and appraisal procedure comparatively ripe at present mainly can be divided into following three classes:

(1) based on the appraisal procedure of reliability theory

The method first to set up the failure model of element, according to failure probability and consequence determination risk indicator.In this approach, the failure state correctly setting up component failure model and selective system is priority application, and the method for current selective system state mainly contains analytical method, Monte Carlo simulation approach and method that both combine.Analytical method (i.e. State enumeration method) is the various malfunctions listing system as much as possible, and one by one selects according to suitable criterion, then the consequence of the selected state of assessment, the risk indicator of last computing system.Monte Carlo simulation approach is a kind of stochastic simulation mathematical method, and first it set up a probability simulation or stochastic process, carried out the running status of simulation system by probability sampling, according to the statistical nature computing system risk indicator of sample.

Can the state of simulate system on physical layer based on the appraisal procedure of reliability theory, and accurate system failure consequence can be drawn by Load flow calculation.But deficiency is the impact being difficult to the uncertain factors such as simulation is truly artificial, disaster, and certain methods exists the contradiction between computational accuracy and computing time.

(2) based on the appraisal procedure of risk management

Based on the integrated evaluating method that quantification and qualification normally combines by the power system security risk assessment of risk management.The process of assessment models and evaluation index (comprising qualitative index and quantitative target) is normally set up in qualitative analysis, require that evaluator is familiar with attribute, the feature of evaluation object (system) very much, and need the impact considering external various factors, the normal method used has fault tree models, risk coordinate diagram etc.Quantitative calculating is normally chosen corresponding real data and is calculated every evaluation index, to some qualitatively index need first by its quantification, finally calculate the numerical value (risk of system) of assessment objective according to indices.Key link wherein determines the weight of each index, and at present the more method of application has levels analytic approach, entropy assessment, factor analysis, Grey System Method etc.

The physical arrangement of the less consideration system of this type of methods of risk assessment, and the impact of subjective factor is subject to when agriculture products weight.But well quantification and qualification can be combined together, be conducive to giving full play to expertise, and be convenient to electric power enterprise in managerial application.

(3) based on the appraisal procedure of the new theories such as artificial intelligence

In recent years, rapidly, various new algorithm is also constantly perfect, and the research based on the Study of Risk Evaluation Analysis for Power System of these new theories, new method also receives the favor of experts and scholars in the new theory development such as artificial intelligence.Such as, Chen Weihua etc. have carried out deep analysis to the application of fuzzy neural network in Study of Risk Evaluation Analysis for Power System in " Risk assessment of cascading failure in power system based on fuzzy neural network " (2007), and establish model and the evaluation index of cascading failure risk assessment on this basis.Ren Hui etc. in " cascading failure in power system venture analysis " (2009) in all its bearings (as complex network) studied the evolution of cascading failure in power system and the model of analysis in great detail, have studied the feasibility adopting branching process model to carry out actual electric network risk management, providing certain directive function for formulating the measure reducing cascading failure risk.Although these new methods are development and improvement theory of risk assessment in some aspects, also has certain distance apart from practical application.The method for supervising at present power communication system used, part index number is too outmoded on the one hand, the change difference of network quality cannot be gone out by real embodiment, the simple monitoring of existing method just to index observed reading on the other hand, further do not excavate its internal information, thus cannot the supportability of overall evaluation system and the potential risk of accurate managing and control system.

Summary of the invention

The technical assignment of the technical problem to be solved in the present invention and proposition carries out improving to prior art and improves, there is provided the various dimensions operation risk evaluation method towards communication network, to reach the accurate early warning ground object of accurate location and the potential risk realizing particular problem in operation.For this reason, the present invention takes following technical scheme.

Towards the various dimensions operation risk evaluation method of communication network, it is characterized in that comprising the following steps:

1) the various dimensions operation risk set up towards communication network is evaluated and Early-warning Model, electric system is divided into four layers, is respectively the base values layer, key element layer, business index layer, the coefficient indices layer that are positioned at bottom;

2) determine evaluation criterion weight, adopt the relative importance that in analytical hierarchy process determination assessment indicator system, index at different levels is run for power communication system, reflect the size of index role in evaluation in quantitative mode;

3) determine evaluation index code of points, base values is classified, be divided into trend to require the index of extremalization, require to be stabilized in index and the qualitative index of a certain ideal value; Trend requires that the index that the index of extremalization and requirement are stabilized in a certain ideal value selects discrete method between linear zone to be the marking mode of index, and qualitative index selects grade percentile method to be the marking mode of index;

4) power communication system comprehensive grading management, color management is carried out to the power communication system comprehensive grading value that the evaluation of various dimensions operation risk and the Early-warning Model towards communication network obtains, to know the different management and control schemes of identification in difference scoring situation, guidance system daily management and maintenance.

As improving further and supplementing technique scheme, the present invention also comprises following additional technical feature.

Base values layer runs mass of foundation according to powerline network, resource distribution, network capabilities, external influence factors set up general evaluation index storehouse, and the key-factor-evaluation for each core business provides and quantizes basis; Key element layer: according to weighted basis index, four aspect key elements of comprehensive evaluation every core business operational support ability: history run index, resource distribution index, network capabilities exponential sum external action index; Business index layer: according to each business key element of weighting, the operation risk index of the every core business of comprehensive evaluation, core business comprises scheduling production business and the large class of management information business two; System index layer: according to each core business value-at-risk of weighting, the overall operation value-at-risk of comprehensive evaluation power communication system, reflects the risk situation of whole power communication system.

Base values layer comprises Service assurance rate, the accredited rate of personnel, independent dual power supply ratio, optical cable coverage rate, passage available rate, personnel depaly rate, Cross Trade protection ratio, website to become in ring rate, archive rate, Unit Assets cost, the network coverage, network broadband utilization factor one or more.

When determining evaluation criterion weight, in secondary about last layer to each factor of same level, the importance of some factors compares between two, and Judgement Matricies, analyzes its relative importance, obtain corresponding proper vector according to its characteristic root, after normalization, obtain the weight vectors of each factor.

When determining evaluation criterion weight, first according to set up base values system hierarchy Model, set up the judgment matrix compared between two.If factor Ak is relevant with lower floor element in last layer time, represent factor Ak with Bij, element Bi is to the relative importance of element B j, and comparative result can arrange into following matrix, i.e. judgment matrix:

When Bij is 1 or 2, for last layer time certain factor, this layer of factor Bi and Bj no less important; When Bij is 3 or 4, for last layer time certain factor, this layer of factor Bi is more important a little than Bj; When Bij is 5 or 6, for last layer time certain factor, this layer of factor Bi is more obvious than Bj important; Work as Bij

When being 7 or 8, for last layer time certain factor, this layer of factor Bi is stronger than Bj important; Work as Bij

When being 9 or 10, for last layer time certain factor, this layer of factor Bi is extremely more important than Bj; After obtaining judgment matrix B, obtain its eigenvalue of maximum λ max, then solve characteristic of correspondence vector according to characteristic of correspondence equation BW=λ maxW, the weight coefficient of each index of same level will be after W normalization; Carry out consistency check to judgment matrix, definition coincident indicator C.I. is: when each judgment matrix C.I.≤0.1, then think that matrix has consistance, accept accordingly and the weight coefficient of calculating, otherwise readjust judgment matrix.

Weight coefficient computing method are:

401) judgment matrix is standardized by row,

$\stackrel{\‾}{B_{\mathrm{ij}}}=\frac{B_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{B}_{\mathrm{ij}}}(i,j=\mathrm{1,2}...n)$

402) be added by row and number

$\stackrel{\‾}{W_i}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{B_{\mathrm{ij}}}$

403) be normalized, obtain weight coefficient W _i,

$W_i=\frac{\stackrel{\‾}{W_i}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{W_i}}.$

Beneficial effect: this programme is in conjunction with power industry feature, by considering networking, the method of operation, operation management, human resources, fund input, the cross influence of external action equivalent risk factor, establish the various dimensions operation risk evaluation towards communication network and Early-warning Model, three grades are carried out to evaluation index and evaluates segmentation and combination weights, novel evaluation algorithms is adopted to carry out science to index system, mark accurately, form the evaluation result of classification refinement, and formulated power communication system comprehensive grading management method, color management is adopted to system synthesis health degree.

By the application of this programme, can not only the dynamic operation quality of each business of quantitatively evaluating and power communication system, more can realize the accurate location of particular problem and the accurate early warning of potential risk in the network operation, for promoting its Service assurance ability and integrated management level further, resource continuous property input and structural allocation is instructed to provide the new type management instrument of precise quantification.Be specially:

1, this programme is from power communication system entirety, comprehensive combing comprises the factor that networking, the method for operation, operation management, human resources, fund input, external action etc. may have an impact to system cloud gray model quality, with systemic, scientific, comprehensive for the overall principle, set up and the perfect assessment indicator system of power communication system supportability, thus form various dimensions operation risk and evaluate and Early-warning Model;

2, this programme has carried out three grades of evaluation segmentations and combination weights to evaluation index, form the evaluation result of classification refinement, by whole evaluation differentiation reflection macroscopic result, and can accurate location particular problem be evaluated by each dimension subitem and Predict is carried out to future developing trend and potential risk;

3, this programme the maintenance resources such as dynamic reflection human, financial, and material resources can drop into and associate the quantification of system cloud gray model risk, the efficiency that can not only drop into existing maintenance resources and benefit carry out quantitatively evaluating, have important directive significance in the total amount simultaneously dropped in resource continuous property and structural allocation;

4, this programme formulates power communication system comprehensive grading management method further on the basis of appraisement system, color management is carried out to system comprehensive evaluation result, and for each color region, corresponding risk management and control strategy is proposed, for the daily management mission of system provides important evidence.

Accompanying drawing explanation

Fig. 1 is various dimensions operation risk of the present invention evaluation and Early-warning Model figure.

Fig. 2 is that the present invention marks process flow diagram.

Embodiment

Below in conjunction with Figure of description, technical scheme of the present invention is described in further detail.

The present invention includes following steps:

One, the various dimensions operation risk evaluation towards communication network and Early-warning Model is set up

In order to science, comprehensively all kinds of factor of evaluation run the impact produced on power communication system, the various dimensions operation risk that this programme is set up towards communication network is evaluated and Early-warning Model, and as shown in Figure 1, model entirety comprises four levels altogether:

1, base values layer: run the aspects such as mass of foundation, resource distribution, network capabilities, external influence factors with powerline network, set up general evaluation index storehouse, the key-factor-evaluation for each core business provides and quantizes basis;

2, key element layer: weighted basis index, four aspect key elements of comprehensive evaluation every core business operational support ability: history run index, resource distribution index, network capabilities exponential sum external action index;

3, business index layer: each business key element of weighting, the operation risk index of the every core business of comprehensive evaluation, core business comprises scheduling production business and the large class of management information business two;

4, system index layer: each core business value-at-risk of weighting, the overall operation value-at-risk of comprehensive evaluation power communication system, reflects the risk situation of whole power communication system.

Risk assessment and Early-warning Model are from power industry feature and practical operation, analyze the dynamic associations of history run, resource distribution, network capabilities and this four dimensions of external action, selected multiple base values cover all factors in the inside and outside affecting power communication system running quality, achieve system science, objective, comprehensive risk assessment, point item rating simultaneously by arranging the weight of each level base values of each dimension and between variable region, can Precise Diagnosis system exist abnormal problem, the potential risk of depth recognition system.This model has stronger science, feasibility, applicability and perspective, refers to the accurate quantification tool of the daily management of conducting communication network and maintenance.

Two, evaluation criterion weight is determined

Determine the weight of evaluation index, namely be the weight of setting evaluation model underlying basis index, namely determining the relative importance that in assessment indicator system, index at different levels is run for power communication system, is the size reflecting index role in evaluation in quantitative mode.

The defining method of current weight has a lot, this programme considers various method relative merits in actual applications, and the complexity of data acquisition and cost, weight for base values is determined to select analytical hierarchy process (AHP method), makes every effort to the relative importance truly reflecting index reality.

Analytical hierarchy process is a kind of multi-level weight coefficient analytical method, and the various decision factor of mode process combined with quantitative and qualitative analysis is the effective ways made a policy to challenge.The ultimate principle of Weight of Coefficient through Analytic Hierarchy Process is utilized to be exactly compare between two about the importance of some factors in last layer time each factor of same level, Judgement Matricies, analyze its relative importance, obtain corresponding proper vector according to its characteristic root, after normalization, obtain the weight vectors of each factor.

First according to set up base values system hierarchy Model, the judgment matrix compared between two is set up.If last layer time middle factor A _kwith lower floor element B={ B ₁, B ¹..., B _nrelevant, represent factor A with Bij _k, element Bi is to the relative importance of element B j, and comparative result can arrange into following matrix, i.e. judgment matrix B=(B _ij) _n*n:

Relative importance degree wherein represented by Bij represents with numerical value in following table.

Table 1 factor relative importance degree

After obtaining judgment matrix B, obtain its eigenvalue of maximum λ max, recycle its characteristic of correspondence equation BW=λ maxW and solve characteristic of correspondence vector W=(w ₁, w ₂..., w _n) ^t, the weight coefficient of each index of same level will be after W normalization.

Can not have crash consistency with comparing the judgment matrix drawn between two, therefore will carry out consistency check to it, definition coincident indicator C.I. is

$C.I.=\frac{{\mathrm{\λ}}_{\max }-n}{n-1}$

Generally, when each judgment matrix C.I.≤0.1, just think that matrix has consistance, accordingly and the weight coefficient calculated is acceptable, otherwise must judgment matrix be adjusted.

Choose eigenvalue of maximum and the proper vector of Sum-Product algorithm approximate treatment judgment matrix in this programme, step is as follows:

1, judgment matrix is standardized by row,

$\stackrel{\‾}{B_{\mathrm{ij}}}=\frac{B_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{B}_{\mathrm{ij}}}(i,j=\mathrm{1,2}...n)$

2, be added by row and number

$\stackrel{\‾}{W_i}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{B_{\mathrm{ij}}}$

3, be normalized, obtain weight coefficient W _i,

$W_i=\frac{\stackrel{\‾}{W_i}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{W_i}}$

Three, evaluation index code of points is determined

Determining the code of points of evaluation index, is namely determine that how the desired value of the base values intuitively observed being converted to dimensionless in a certain interval can be used for comparing the process of the score value of calculating.In this programme, concrete scoring process as shown in Figure 2.

Because this programme index system is huge, for numerous evaluation indexes, according to its characteristic, index is divided into following three classes, and selects different methods of marking respectively.

1, trend requires the index of extremalization: according to the object of project evaluation power communication system health degree, one class is benefit property index, the i.e. index that is the bigger the better of score value, as indexs such as Service assurance rate, passage available rates, another kind of is profit and loss index, namely the index that score value is the smaller the better, as indexs such as flow of personnel rate, hundred metanetwork assets protections always drop into;

2, require to be stabilized in the index of a certain ideal value: the score value of this type of index and the health degree of power communication system disproportionate, its score value should remain on a certain reasonable interval, too high or too low all can the running quality of influential system, as indexs such as personnel's mean age, personnel length of service;

3, qualitative index: this type of index directly cannot calculate score value by data, and need formulate corresponding evaluation criterion to evaluation index, reflects its evaluation result, as the index such as power load, urban construction with objective specification and analysis.

Consider the complexity of index properties and practical operation, for first two index, select discrete method between linear zone to be the marking mode of index.Between linear zone, discrete method carries out rank to desired value, and actual optimal parameter value and the poorest desired value are chosen as best result and minimum point respectively, and other index score values are evenly distributed in the interval of best result and minimum point.The method guarantees that scoring is all in a certain interval, increases the gap amplitude (discreteness) of scoring, makes evaluation difference domination more.

Between use linear zone before discrete method, the trend requirement of first unified each index, according to project evaluation object, is the bigger the better as criterion with index score value, eliminates the incommensurability between index.By corresponding 100 points of optimal parameter value, corresponding 50 points of the poorest desired value (best result and minimum point can need to carry out flexible according to evaluation), different indexs is required for trend, (wherein x is desired value all to utilize basic formula y=ax+b, y is this index score value, a, b are relevant parameter), base values score value is evenly distributed in 50 and assigns in 100 by stages, form normalization evaluation result.Such as index " Service assurance rate ", get a mark of 100 when definition desired value is 100%, obtain 50 points when 95%, can calculate conversion formula is y=10x-900, and when Service assurance rate is 98%, scoring is 80 points.

Cannot directly by index that data calculate for the third, grade percentile method is selected to be the marking mode of index, namely according to actual conditions, each desired value is divided into some levels or shelves (generally dividing 3 ~ 5 grades) from low to high by order of quality, every grade of given score value, actual desired value according to partition of the level standard determination rank belonging to it, thus obtains corresponding score value.Such as:

Table 2 grade percentile method example

After being marked to all base values by above two kinds of methods of marking, by all index score values according to formula (wherein n is index number, and x is index mark, and w is corresponding index weights) is weighted summation, can obtain the comprehensive grading value of whole system.

Four, power communication system comprehensive grading management

Color management is carried out to the power communication system comprehensive grading value that application operation risk evaluation and Early-warning Model obtain, proposes the different management and control schemes of system in difference scoring situation, as shown in table 3, guidance system daily management and maintenance.

The color management of table 3 comprehensive grading

More than for the various dimensions operation risk evaluation method towards communication network shown in Fig. 1,2 is specific embodiments of the invention; substantive distinguishing features of the present invention and progress are embodied; can according to the use needs of reality; under enlightenment of the present invention; it is carried out to the equivalent modifications of the aspect such as shape, structure, all at the row of the protection domain of this programme.

Claims (6)

1., towards the various dimensions operation risk evaluation method of communication network, it is characterized in that comprising the following steps:

1) the various dimensions operation risk set up towards communication network is evaluated and Early-warning Model, electric system is divided into four layers, is respectively the base values layer, key element layer, business index layer, the coefficient indices layer that are positioned at bottom;

2) determine evaluation criterion weight, adopt the relative importance that in analytical hierarchy process determination assessment indicator system, index at different levels is run for power communication system, reflect the size of index role in evaluation in quantitative mode;

3) determine evaluation index code of points, base values is classified, be divided into trend to require the index of extremalization, require to be stabilized in index and the qualitative index of a certain ideal value; Trend requires that the index that the index of extremalization and requirement are stabilized in a certain ideal value selects discrete method between linear zone to be the marking mode of index, and qualitative index selects grade percentile method to be the marking mode of index;

4) power communication system comprehensive grading management, color management is carried out to the power communication system comprehensive grading value that the evaluation of various dimensions operation risk and the Early-warning Model towards communication network obtains, to know the different management and control schemes of identification in difference scoring situation, guidance system daily management and maintenance.

2. the various dimensions operation risk evaluation method towards communication network according to claim 1, it is characterized in that: base values layer runs mass of foundation according to powerline network, resource distribution, network capabilities, external influence factors set up general evaluation index storehouse, and the key-factor-evaluation for each core business provides and quantizes basis; Key element layer: according to weighted basis index, four aspect key elements of comprehensive evaluation every core business operational support ability: history run index, resource distribution index, network capabilities exponential sum external action index; Business index layer: according to each business key element of weighting, the operation risk index of the every core business of comprehensive evaluation, core business comprises scheduling production business and the large class of management information business two; System index layer: according to each core business value-at-risk of weighting, the overall operation value-at-risk of comprehensive evaluation power communication system, reflects the risk situation of whole power communication system.

3. the various dimensions operation risk evaluation method towards communication network according to claim 2, is characterized in that: base values layer comprises Service assurance rate, the accredited rate of personnel, independent dual power supply ratio, optical cable coverage rate, passage available rate, personnel depaly rate, Cross Trade protection ratio, website to become in ring rate, archive rate, Unit Assets cost, the network coverage, network broadband utilization factor one or more.

4. the various dimensions operation risk evaluation method towards communication network according to claim 1, it is characterized in that: when determining evaluation criterion weight, in secondary about last layer to each factor of same level, the importance of some factors compares between two, Judgement Matricies, analyze its relative importance, obtain corresponding proper vector according to its characteristic root, after normalization, obtain the weight vectors of each factor.

5. the various dimensions operation risk evaluation method towards communication network according to claim 4, is characterized in that: when determining evaluation criterion weight, first according to set up base values system hierarchy Model, sets up the judgment matrix compared between two.If factor Ak is relevant with lower floor element in last layer time, represent factor Ak with Bij, element Bi is to the relative importance of element B j, and comparative result can arrange into following matrix, i.e. judgment matrix:

When Bij is 1 or 2, for last layer time certain factor, this layer of factor Bi and Bj no less important; When Bij is 3 or 4, for last layer time certain factor, this layer of factor Bi is more important a little than Bj; When Bij is 5 or 6, for last layer time certain factor, this layer of factor Bi is more obvious than Bj important; When Bij is 7 or 8, for last layer time certain factor, this layer of factor Bi is stronger than Bj important; When Bij is 9 or 10, for last layer time certain factor, this layer of factor Bi is extremely more important than Bj; After obtaining judgment matrix B, obtain its eigenvalue of maximum λ max, then solve characteristic of correspondence vector according to characteristic of correspondence equation BW=λ maxW, the weight coefficient of each index of same level will be after W normalization; Carry out consistency check to judgment matrix, definition coincident indicator C.I. is: when each judgment matrix C.I.≤0.1, then think that matrix has consistance, accept accordingly and the weight coefficient of calculating, otherwise readjust judgment matrix.

6. the various dimensions operation risk evaluation method towards communication network according to claim 5, is characterized in that: weight coefficient computing method are:

401) judgment matrix is standardized by row,

$\stackrel{\‾}{B_{\mathrm{ij}}}=\frac{B_{\mathrm{ij}}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{B}_{\mathrm{ij}}}(i,j=\mathrm{1,2}...n)$

402) be added by row and number

$\stackrel{\‾}{W_i}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{B_{\mathrm{ij}}}$

403) be normalized, obtain weight coefficient W _i,

$W_i=\frac{\stackrel{\‾}{W_i}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\stackrel{\‾}{W_i}}$ 。