CN104392391A - Power grid running safety risk quantification method - Google Patents

Abstract

The invention provides a power grid running safety risk quantification method. The method comprises establishing a risk quantification model, wherein the power grid value risk is equal to consequence value*probability value; determining the category of influence factor items required to participate in risk quantification estimation to obtain the weighted value among all influence factor sub-items contained by each screened influence factor item; selecting the corresponding influence factor sub-items from each influence factor items participating in risk quantification estimation, multiplying the selected influence factor sub-items to obtain a probability value, multiplying the probability value by a consequence value to obtain the power grid risk value during the occurring of the dangerous factors of the power grid, and obtaining the power grid grade according to the power grid risk value. By the aid of the method, high risk quantification scientificity and accuracy is provided, and risks can be quantified reasonably to be used as the object measurement basis for guiding actual power generation.

Description

A kind of safe operation of electric network risk quantification method

Technical field

The invention belongs to power grid risk quantification technique field, be specifically related to a kind of safe operation of electric network risk quantification method.

Background technology

In recent years, China's power network development is rapid, and the requirement of all trades and professions to power network safety operation of society is more and more higher, electrical network once break down, the economic loss caused and social influence increasing.Therefore, effective risk assessment is carried out to electrical network, that is: the risk assessment based on hazard factor existed in a period of time of being caused by reasons such as equipment stoppage in transit, platform test, unit exceptions; Then, on the basis of risk assessment, constituent parts formulates the power grid risk control measure of administrative area, comprise technical measures and organiztion measure etc., reduce the possibility that risk causes severity of consequence and risk to occur, by risk control in acceptable degree, have important practical significance.

In prior art, the main quilitative method that adopts carries out problem-targeted risk assessment to electrical network, has the problems such as risk evaluation result out of true, subjectivity are strong, is difficult to the demand that satisfied current relevant unit assesses power grid risk.

Summary of the invention

For the defect that prior art exists, the invention provides a kind of safe operation of electric network risk quantification method, can effectively solve the problem.

The technical solution used in the present invention is as follows:

The invention provides a kind of safe operation of electric network risk quantification method, comprise the following steps:

S1, sets up risk quantification model, and the risk quantification algorithm that described risk quantification model adopts, for connecting multiplication algorithm, obtains the power grid risk value universal calculation equation shown in formula one thus:

Power grid risk value=consequence value * probable value; Formula one

Wherein, consequence value is the value that the consequence brought after being occurred by power grid risk carries out quantizing, and its expression formula is: consequence value=consequence is seriously worth * social influence factor; Wherein, consequence be seriously worth refer to the loss load of evaluated electrical network to carry out quantification conversion after obtain numerical value; The value of social influence factor and the guarantor of evaluated electrical network power rank positive correlation, and it is higher to protect power supply rank, then the value of social influence factor is larger;

Probable value=factor to affect item A1* factor to affect item A2 ... factor to affect item Ab; B is the sum screening the factor to affect item obtained;

S2, the grid equipment fault database that described risk quantification model read is up-to-date, obtains some fault cases between a certain historical years T1 to current year T2, and each fault case records m1 bar factor to affect item when there is certain fault; Wherein, each factor to affect item includes several factor to affect subitems;

S3, determines the classification needing the factor to affect item participating in risk quantification assessment:

Be specially:

S3.1, described risk quantification model calculates the information entropy of every bar factor to affect item in m1 bar factor to affect item respectively;

S3.2, described risk quantification model sorts to m1 bar factor to affect item by information entropy order from high to low; Herein, information entropy reacts the significance level of corresponding factor to affect item; Then, the m2 bar factor to affect item be arranged in front is filtered out; Wherein, 1≤m2≤m1, makes the summation of the information entropy of filtered out m2 bar factor to affect item reach the number percent of set information amount; Described m2 bar factor to affect item is the factor to affect item of the participation risk quantification assessment filtered out;

S4, for filtered out each factor to affect item, it forms by several factor to affect subitems; Obtain the weighted value between each factor to affect subitem that each described factor to affect item comprises;

Wherein, for device type factor to affect item, the weighted value between the factor to affect subitem that it comprises obtains by the following method:

S4.1, supposes that described device type factor to affect item comprises z device type factor to affect subitem;

S4.2, reads the annual fault statistics table of electrical network and third party's fault statistics table; Wherein, the number of times of each device type factor to affect subitem generation forced outage corresponding to described electrical network annual fault statistics table record different year; The forced outage rate of each device type factor to affect subitem corresponding to described third party's fault statistics table record different year;

S4.3, obtain the record between historical years T1 to current year T2 in described electrical network year fault statistics table, then, calculate the average time that forced outage occurs often kind of device type factor to affect subitem between described T1 to T2 year, be designated as C1, C2 respectively ... Cz;

Obtain the record between historical years T1 to current year T2 in described third party's fault statistics table, then, calculate the mean value of often kind of device type factor to affect subitem forced outage rate between described T1 to T2 year, be designated as D1, D2 respectively ... Dz;

S4.4, if the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: fault mean value=(Ci*Q1+Di*Q2)/2 of often kind of device type factor to affect subitem;

S4.5, is normalized the fault mean value of calculated various device type factor to affect subitems, obtains the weighted value of various device type factor to affect subitem;

Described historical statistics factor to affect item is obtained by following formulae discovery:

Value=every the annual of 1+ same category of device of historical statistics factor to affect item is broken down the mean value of number of times/same category of device sum=1+ same category of device failure rate over the years;

S5, after the weighted value between each factor to affect subitem that the factor to affect item classification and often kind of factor to affect item that obtain participating in risk quantification assessment comprise; The electrical network hazard factor of quantitative evaluation as required, corresponding factor to affect subitem is chosen from the factor to affect item of each participation risk quantification assessment, each selected factor to affect subitem is multiplied and namely obtains probable value, again probable value is multiplied by consequence value, namely obtains power grid risk value when described hazard factor occurs electrical network;

S6, the corresponding relation of predefine power grid risk value scope and power grid risk grade; By searching described corresponding relation, obtain the power grid risk grade corresponding to power grid risk value that S5 obtains; And show described power grid risk grade by patterned way.

Preferably, consequence is seriously worth value by the following method:

The security incident grade of standard comprises 5 ranks, is respectively: especially big loss level, heavy losses rank, greater loss rank, general category-A loss level and general category-B loss level; The span that consequence corresponding to each rank is seriously worth sees the following form:

Rank	Consequence is seriously worth value
		Especially big loss	More than 3000
Heavy losses	500 < consequences are seriously worth≤and 3000
		Greater loss	100 < consequences are seriously worth≤and 500
General category-A loss	30 < consequences are seriously worth≤and 100
		General category-B loss	1 < consequence is seriously worth≤and 30

Obtain consequence thus to be seriously worth.

Preferably, social influence because of the several social influence factor subitem comprised for superfine protecting power supply factor to affect subitem, one-level protects power supply factor to affect subitem, secondary protects power supply factor to affect subitem, particular time protects power supply factor to affect subitem and general period factor to affect subitem; The value of each factor to affect subitem is:

Period	Social influence factor value
		Superfine guarantor is powered	2
One-level protects power supply	1.6
		Secondary protects power supply	1.4
Particular time protects power supply	1.2
		General period	1

Obtain the value of each factor to affect subitem in social influence factor thus.

Preferably, S3 is specially:

Information entropy by following formulae discovery every bar factor to affect item:

$H_s=-\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{p}_i{\log }_2{p}_i$

Wherein, H _sfor the information entropy of factor to affect item, p _ifor the probability of malfunction of factor to affect subitem between a certain historical juncture T1 to current time T2; The sum of the factor to affect subitem that n comprises for factor to affect item.

Preferably, for the other types factor to affect item except device type factor to affect item and historical statistics factor to affect item participating in risk quantification assessment, referred to as factor to affect item S, then: for factor to affect item S, the weighted value between its each factor to affect subitem is obtained by following steps S4-6:

S4-6, in each factor to affect subitem that factor to affect item S comprises, chooses any one factor to affect subitem as baseline impact factor subitem, other influences factor subitem is called non-referenced factor to affect subitem;

Read the electrical network year fault statistics table between historical years T1 to current year T2, count baseline impact factor subitem occurrence number and total effective degree, draw baseline impact factor subitem probability of occurrence; Count non-referenced factor to affect subitem occurrence number and total effective degree again, draw non-referenced factor to affect subitem probability of occurrence; By each non-referenced factor to affect subitem probability of occurrence divided by baseline impact factor subitem probability of occurrence, obtain normalized non-referenced factor to affect subitem weight K1, wherein, baseline impact factor subitem weight is 1;

Read the third party's fault statistics table between historical years T1 to current year T2, calculate the year number of stoppages average value P 1 of often kind of non-referenced factor to affect subitem, and the year number of stoppages average value P 2, P1 of going back Calculation Basis factor to affect subitem obtains weight K2 after normalization divided by P2;

Again because the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: weighted value=(K1*Q1+K2*Q2)/2 of often kind of non-referenced factor to affect subitem.

Preferably, after calculating the weighted value in factor to affect item S between each factor to affect subitem, also comprise: the step that calculated weighted value is verified; If met with reference to span, or in the predetermined amplitude of reference value, then calculated weighted value is the weighted value finally calculated; Otherwise, expand the annual scope participating in the annual fault statistics table of electrical network and third party's fault statistics table adding up, and re-execute S4-6, until meet;

Concrete, factor to affect item S comprises site operation because several, control measure are because several, fault category is because several, weather effect is because several, equipment deficiency is because several, overhaul management is because several and repair time is because of the one or more combination in several;

Reference span or the reference value of the weighted value between each factor to affect subitem that factor to affect item S comprises are respectively:

Site operation comprises two factor to affect subitems because of several, is respectively: site environment does not use large-scale repair apparatus factor to affect subitem and site environment to use large-scale repair apparatus factor to affect subitem; Site environment does not use the reference value of large-scale repair apparatus factor to affect subitem to be 1; Site environment uses the span of large-scale repair apparatus factor to affect subitem to be: 1 < site environment uses span≤2 of large-scale repair apparatus factor to affect subitem;

Control measure comprise two factor to affect subitems because of several, are respectively: adopt spy's equipment of patrolling to carry out patrolling and examining and do not adopt spy's equipment of patrolling to patrol and examine; Wherein, when adopting the spy equipment of patrolling to patrol and examine, the probability of electrical network occurrence risk diminishes, then the value that the special equipment of patrolling carries out patrolling and examining is less than and does not adopt spy's equipment of patrolling to carry out the value of patrolling and examining; Further, 0< control measure are because of each factor to affect subitem span≤1 in several;

Fault category is because of several: bear the requirement determination fault category of the safety and stability standard of maximum perturbation ability according to power platform because of several values; Concrete, one-level, secondary and three level security stabilized references are corresponding first kind fault, Equations of The Second Kind fault and the 3rd class fault respectively; That is, fault category comprises three fault category factor to affect subitems because of several, and the value of each fault category factor to affect subitem sees the following form:

Fault category

Fault category factor to affect subitem value

First kind fault	1
		Equations of The Second Kind fault	0.2～0.6
3rd class fault	0～0.2

Weather effect is because of several: according to the weather condition of maintenance time in power grid maintenance scheme, and determine that weather effect is because of several values, weather condition is more severe, and the probability of electrical network occurrence risk is higher, then weather effect because of several values larger; Weather effect sees the following form because of the concrete value of the several factor to affect subitem comprised:

Weather pattern (weather effect factor subitem)	The value of weather effect factor subitem
		Normally	1
Thunderstorm gale	1～2
		Risk of forest fire	1～1.5
High temperature	1～1.2
		Dense fog	1～1.2
Heavy snow	1～4
		Freeze	1～3

Equipment deficiency is because of several: according to the serious conditions of equipment maintained in power grid maintenance scheme whether existing defects and defect, and determine that equipment deficiency is because of several concrete values; Concrete, equipment deficiency is more serious, and the probability of generation power grid risk is higher, then equipment deficiency because of several values larger; Therefore, equipment deficiency sees the following form because of the concrete value of the several each factor to affect subitem comprised:

Overhaul management is because of several: overhaul type according in power grid maintenance scheme, and determine that overhaul management is because of several values, wherein, scheduled overhaul can not produce risk to electrical network, and its value is 1; And all there is possibility electrical network being brought to risk in non-plan repair and accident maintenance; That is, overhaul management because of the reference span of the several each factor to affect subitem comprised as follows:

Maintenance type (that is: overhaul management factor to affect subitem)	Overhaul management factor to affect subitem value
		Scheduled overhaul	1
Non-plan repair	1.5
		Accident maintenance	2

Repair time is because of several: determine that the repair time is because of several values according to repair rate in power grid maintenance scheme, wherein, repair rate is higher, then the probability of electrical network occurrence risk is lower, then the repair time because of several values less; That is, the repair time sees the following form because of the concrete value of reference of the several each factor to affect subitem comprised:

Repair rate (that is: repair time factor to affect subitem)	Repair time factor to affect subitem value
		1 ~ 3 day	0.3～0.5
3 ~ 10 days	0.5～1
		10 ~ 30 days	1～1.5
More than 30 days	1.5～3

Reference value or the reference span of each factor to affect subitem that factor to affect S comprises can be obtained thus.

Preferably, also comprise:

S7, sets up preliminary classification device, using many groups corresponding relation of known hazard factor and power grid risk grade as training sample, uses described training sample to train described preliminary classification device, constantly the parameter of the described preliminary classification device of adjustment; Thus obtain the sorter that trains;

S8, choose different from described training sample many groups and verify sample, the input of the sorter using the hazard factor in described checking sample to obtain as S7, judge that whether the output of sorter is consistent with the power grid risk grade in checking sample, if consistent, then show that the sorter that S7 obtains passes through checking, perform

S8; Otherwise, continue to choose other training samples, the sorter that S7 obtains trained further, until obtain the sorter by checking;

S9, when needs carry out risk class division to a certain hazard factor, directly described hazard factor is input to the described sorter by checking, described sorter directly exports final risk class.

Preferably, the sorting algorithm that described sorter adopts is the combination of one or more in decision Tree algorithms, neural network algorithm, Bayes net algorithm, genetic algorithm and algorithm of support vector machine.

Beneficial effect of the present invention is as follows:

Safe operation of electric network risk quantification method provided by the invention, has science and the accuracy of higher risk quantification assessment, reasonably can quantize, as the objective measure foundation instructing electric power actual production to risk.

Accompanying drawing explanation

Fig. 1 is the schematic flow sheet of safe operation of electric network risk quantification method provided by the invention.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in detail:

As shown in Figure 1, the invention provides a kind of safe operation of electric network risk quantification method, comprise the following steps:

S1, sets up risk quantification model;

The risk assessment link that risk quantification model is risk Whole Course Management provides assessment foundation, and risk quantification model needs the algorithm of a set of specialty and data to be that risk quantification platform does technical support.Wherein, risk quantification algorithm adopts continued multiplication model, and the annual fault statistics table of the grid equipment fault database of data acquisition electrical network, electrical network and third party's fault statistics table, carry out the basis of risk quantification as risk quantification model.

In the present invention, the risk quantification algorithm that described risk quantification model adopts, for connecting multiplication algorithm, obtains the power grid risk value universal calculation equation shown in formula one thus:

Power grid risk value=consequence value * probable value; Formula one

Below the calculating of consequence value and probable value is introduced respectively:

(1) consequence value

Consequence value is the value that the consequence brought after being occurred by power grid risk carries out quantizing, and its expression formula is: consequence value=consequence is seriously worth * social influence factor.

(1) consequence is seriously worth value

It is the norm of electrical network under some dimensions such as loss load, electric voltage exception, frequency anomaly, overload, loss of stability, off-the-line that consequence is seriously worth essence, and calculate for simplifying, the present invention obtains consequence after the loss load of correspondence being carried out quantification conversion and is seriously worth.Such as, special major accident standard in State Council 599 command is: regional grid subtracts for load more than 30%, supposes that the total load of certain electrical network is 10,000 ten thousand kilowatts, and it subtracts for load is 3,000 ten thousand kilowatts, then consequence is seriously worth value is more than 3000, and other standards by that analogy.

In order to grade classification is clearer, the level Four security incident grading standard of State Council 599 command is carried out refinement again, General Loss is divided into the loss of general category-A and the loss of general category-B.The span that the consequence that then each rank is corresponding is seriously worth sees the following form:

Rank	Consequence is seriously worth value
		Especially big loss	More than 3000
Heavy losses	500 < consequences are seriously worth≤and 3000
		Greater loss	100 < consequences are seriously worth≤and 500
General category-A loss	30 < consequences are seriously worth≤and 100
		General category-B loss	1 < consequence is seriously worth≤and 30

Obtain consequence thus to be seriously worth.

(2) value of social influence factor

The value of social influence factor and the guarantor of evaluated electrical network power rank positive correlation, and it is higher to protect power supply rank, then the value of social influence factor is larger;

Social influence because of the several social influence factor subitem comprised for superfine protecting power supply factor to affect subitem, one-level protects power supply factor to affect subitem, secondary protects power supply factor to affect subitem, particular time protects power supply factor to affect subitem and general period factor to affect subitem; The value of each factor to affect subitem is:

Period	Social influence factor value
		Superfine guarantor is powered	2
One-level protects power supply	1.6
		Secondary protects power supply	1.4
Particular time protects power supply	1.2
		General period	1

Obtain the value of each factor to affect subitem in social influence factor thus.

(2) calculating of probable value

Probable value is the value after a series of factor to affect item is multiplied, and its general expression is: probable value=factor to affect item A1* factor to affect item A2 ... factor to affect item Ab; B is the sum screening the factor to affect item obtained;

Device type factor, fault category factor, historical statistics factor, weather factor, equipment deficiency factor, overhaul management factor, repair time factor, site operation factor and control measure factor is obtained finally to screen, then probable value=(device type factor * fault category factor * historical statistics factor * weather factor * equipment deficiency factor * overhaul management factor * repair time factor * site operation factor * control measure factor), introduce the choosing method of factor to affect item and each factor to affect obtaining value method below in detail:

S2, the grid equipment fault database that described risk quantification model read is up-to-date, obtains some fault cases between a certain historical years T1 to current year T2, and each fault case records m1 bar factor to affect item when there is certain fault; Wherein, each factor to affect item includes several factor to affect subitems;

Such as, risk quantification model reads grid equipment fault database by interface, whether the fault pertinency factor according to equipment is added up, and what come out has with risk pertinency factor: device type, fault category, weather effect, equipment deficiency, overhaul management, repair time, site operation, control measure, electric pressure, circuit grade, line length, unit capacity, GIS etc.Then, the factor to affect item that the mode determination finishing screen of the information entropy introduced by subsequent step is elected.

S3, determines the classification needing to participate in the factor to affect item that risk assessment quantizes:

Be specially:

S3.1, described risk quantification model calculates the information entropy of every bar factor to affect item in m1 bar factor to affect item respectively;

Concrete, the information entropy by following formulae discovery every bar factor to affect item:

$H_s=-\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{p}_i{\log }_2{p}_i$

Wherein, H _sfor the information entropy of factor to affect item, p _ifor the probability of malfunction of factor to affect subitem between a certain historical juncture T1 to current time T2; The sum of the factor to affect subitem that n comprises for factor to affect item.

S3.2, described risk quantification model sorts to m1 bar factor to affect item by information entropy order from high to low; Herein, information entropy reacts the significance level of corresponding factor to affect item; Then, the m2 bar factor to affect item be arranged in front is filtered out; Wherein, 1≤m2≤m1, makes the summation of the information entropy of filtered out m2 bar factor to affect item reach the number percent of set information amount; Described m2 bar factor to affect item is the factor to affect item of the participation risk quantification assessment filtered out;

Enumerate a concrete example below:

As following table, be 10 fault cases that the selected time reads:

Have five factor to affect items, the information entropy of each factor to affect item correspondence is respectively:

Device type: P (circuit)=0.3, P (main transformer)=0.2, P (bus)=0.1, P (cable)=0.1, P (generator)=0.1, P (communication power supply)=0.1, H (device type)=2.3142464.

Fault category: P (first kind fault)=0.6, P (Equations of The Second Kind fault)=0.3, P (the 3rd class fault)=0.1, H (fault category)=1.2954619.

Electric pressure: P (220kV)=0.7, P (110kV)=0.2, P (500kV)=0.1, H (electric pressure)=1.1567796.

Weather effect: P (normally)=0.6, P (mountain fire)=0.2, P (typhoon)=0.1, P (thunderstorm gale)=0.1, H (weather effect)=1.5709505.

Overhaul of the equipments manages: P (scheduled overhaul)=0.5, P (non-plan repair)=0.5, H (overhaul of the equipments management)=1.

Its information entropy summation=H (device type)+H (fault category)+H (electric pressure)+H (weather effect)+H (overhaul of the equipments management)=7.3374384, according to each factor information entropy size divided by information entropy summation, draw the number percent of each factor to affect item quantity of information, sorting successively according to importance is: device type, weather effect, fault category, electric pressure, overhaul of the equipments management.

Device type: 31.54%

Weather effect: 21.41%

Fault category: 17.66%

Electric pressure: 15.77%

Overhaul of the equipments manages: 13.63%

According to the proportionate relationship of each factor above, device type, weather effect, fault category, electric pressure 4 kinds of factors just provide the quantity of information of 86.37%.When factor type is more, the quantity of information selecting part factor to affect just can provide 95% can be reached, the even quantity of information of 99%, the factor to affect item be left like this and screened fall.

S4, for filtered out each factor to affect item, it forms by several factor to affect subitems; Obtain the weighted value between each factor to affect subitem that each described factor to affect item comprises;

(1) the weighted value computing method between the factor to affect subitem that comprises of device type factor to affect item

Wherein, for device type factor to affect item, the weighted value between the factor to affect subitem that it comprises obtains by the following method:

S4.1, supposes that described device type factor to affect item comprises z device type factor to affect subitem;

S4.2, reads the annual fault statistics table of electrical network and third party's fault statistics table; Wherein, the number of times of each device type factor to affect subitem generation forced outage corresponding to described electrical network annual fault statistics table record different year; The forced outage rate of each device type factor to affect subitem corresponding to described third party's fault statistics table record different year;

S4.3, obtain the record between historical years T1 to current year T2 in described electrical network year fault statistics table, then, calculate the average time that forced outage occurs often kind of device type factor to affect subitem between described T1 to T2 year, be designated as C1, C2 respectively ... Cz;

Obtain the record between historical years T1 to current year T2 in described third party's fault statistics table, then, calculate the mean value of often kind of device type factor to affect subitem forced outage rate between described T1 to T2 year, be designated as D1, D2 respectively ... Dz;

S4.4, if the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: fault mean value=(Ci*Q1+Di*Q2)/2 of often kind of device type factor to affect subitem;

S4.5, is normalized the fault mean value of calculated various device type factor to affect subitems, obtains the weighted value of various device type factor to affect subitem;

Such as, circuit and unit two factor to affect subitems are comprised for device type factor to affect:

The annual fault statistics table of electrical network is as follows:

Time	Circuit (secondary/hundred kilometer year)	Unit (secondary/platform year)
			2008	1.25	1.79
2009	0.79	0.97
			2010	1.27	0.95
Mean value	1.1	1.23

Third party's fault statistics table is as follows:

Time	Circuit forced outage rate	Unit forced outage rate
			2008	0.437	0.508
2009	0.220	0.333
			2010	0.253	0.518
Mean value	0.303	0.453

Suppose that electrical network annual fault statistics table weight is 80%, third party's fault statistics table weight is 20%.Be calculated as follows:

Line fault mean value=(1.1*0.8+0.303*0.2)/2=0.47,

Unit fault mean value=(1.23*0.8+0.453*0.2)/2=0.54;

Circuit normalization processed, 0.54*1/0.47 ≈ 1.14, that is, when circuit weight is 1, the weight of unit is 1.14.

(2) historical statistics factor to affect item computing method

Described historical statistics factor to affect item is obtained by following formulae discovery:

Value=every the annual of 1+ same category of device of historical statistics factor to affect item is broken down the mean value of number of times/same category of device sum=1+ same category of device failure rate over the years;

Such as, as shown in the table:

Then historical statistics factor value=1.16 of bus.Under normal circumstances, 1≤historical statistics factor span < 2.

The computing method that concrete, in the present invention, historical statistical data factor adopts " the every annual of 1+ same category of device break down number of times/same category of device sum ".Historical statistics factor is for area power grid operation characteristic, to the correction of device type factor.Therefore, historical statistics factor makes it to increase from 1.

(3) the weighted value computing method between the factor to affect subitem that comprises of factor to affect item S

For the other types factor to affect item except device type factor to affect item and historical statistics factor to affect item participating in risk quantification assessment, referred to as factor to affect item S, then: for factor to affect item S, the weighted value between its each factor to affect subitem is obtained by following steps S4-6:

S4-6, in each factor to affect subitem that factor to affect item S comprises, chooses any one factor to affect subitem as baseline impact factor subitem, other influences factor subitem is called non-referenced factor to affect subitem;

Read the electrical network year fault statistics table between historical years T1 to current year T2, count baseline impact factor subitem occurrence number and total effective degree, draw baseline impact factor subitem probability of occurrence; Count non-referenced factor to affect subitem occurrence number and total effective degree again, draw non-referenced factor to affect subitem probability of occurrence; By each non-referenced factor to affect subitem probability of occurrence divided by baseline impact factor subitem probability of occurrence, obtain normalized non-referenced factor to affect subitem weight K1, wherein, baseline impact factor subitem weight is 1;

Read the third party's fault statistics table between historical years T1 to current year T2, calculate the year number of stoppages average value P 1 of often kind of non-referenced factor to affect subitem, and the year number of stoppages average value P 2, P1 of going back Calculation Basis factor to affect subitem obtains weight K2 after normalization divided by P2;

Again because the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: weighted value=(K1*Q1+K2*Q2)/2 of often kind of non-referenced factor to affect subitem.

After calculating the weighted value in factor to affect item S between each factor to affect subitem, also comprise: the step that calculated weighted value is verified; If met with reference to span, or in the predetermined amplitude of reference value, then calculated weighted value is the weighted value finally calculated; Otherwise, expand the annual scope participating in the annual fault statistics table of electrical network and third party's fault statistics table adding up, and re-execute S4-6, until meet;

Concrete, factor to affect item S comprises site operation because several, control measure are because several, fault category is because several, weather effect is because several, equipment deficiency is because several, overhaul management is because several and repair time is because of the one or more combination in several;

Reference span or the reference value of the weighted value between each factor to affect subitem that factor to affect item S comprises are respectively:

Site operation comprises two factor to affect subitems because of several, is respectively: site environment does not use large-scale repair apparatus factor to affect subitem and site environment to use large-scale repair apparatus factor to affect subitem; Site environment does not use the reference value of large-scale repair apparatus factor to affect subitem to be 1; Site environment uses the span of large-scale repair apparatus factor to affect subitem to be: 1 < site environment uses span≤2 of large-scale repair apparatus factor to affect subitem;

Control measure comprise two factor to affect subitems because of several, are respectively: adopt spy's equipment of patrolling to carry out patrolling and examining and do not adopt spy's equipment of patrolling to patrol and examine; Wherein, when adopting the spy equipment of patrolling to patrol and examine, the probability of electrical network occurrence risk diminishes, then the value that the special equipment of patrolling carries out patrolling and examining is less than and does not adopt spy's equipment of patrolling to carry out the value of patrolling and examining; Further, 0< control measure are because of each factor to affect subitem span≤1 in several;

Fault category is because of several: bear the requirement determination fault category of the safety and stability standard of maximum perturbation ability according to power platform because of several values; Concrete, one-level, secondary and three level security stabilized references are corresponding first kind fault, Equations of The Second Kind fault and the 3rd class fault respectively; That is, fault category comprises three fault category factor to affect subitems because of several, and the value of each fault category factor to affect subitem sees the following form:

Fault category	Fault category factor to affect subitem value
		First kind fault	1
Equations of The Second Kind fault	0.2～0.6
		3rd class fault	0～0.2

Weather effect is because of several: according to the weather condition of maintenance time in power grid maintenance scheme, and determine that weather effect is because of several values, weather condition is more severe, and the probability of electrical network occurrence risk is higher, then weather effect because of several values larger; Weather effect sees the following form because of the concrete value of the several factor to affect subitem comprised:

Weather pattern (weather effect factor subitem)	The value of weather effect factor subitem
		Normally	1
Thunderstorm gale	1～2
		Risk of forest fire	1～1.5

High temperature	1～1.2
		Dense fog	1～1.2
Heavy snow	1～4
		Freeze	1～3

typhoon: orange early warning gets 1 ~ 1.2, red early warning gets 1.3 ~ 1.5;

thunderstorm gale: yellow early warning gets 1 ~ 1.2, orange early warning gets 1.3 ~ 1.5, red early warning gets 1.6 ~ 2;

risk of forest fire: orange early warning gets 1 ~ 1.2, and red early warning gets 1.3 ~ 1.5;

high temperature: orange early warning gets 1.1, and red early warning gets 1.2;

dense fog: orange early warning gets 1.1, and red early warning gets 1.2;

heavy snow: yellow early warning gets 1 ~ 1.5, orange early warning gets 1.6 ~ 2.4, and red early warning gets 2.5 ~ 4

freeze: look weather condition and line ice coating situation value.

Equipment deficiency is because of several: according to the serious conditions of equipment maintained in power grid maintenance scheme whether existing defects and defect, and determine that equipment deficiency is because of several concrete values; Concrete, equipment deficiency is more serious, and the probability of generation power grid risk is higher, then equipment deficiency because of several values larger; Therefore, equipment deficiency sees the following form because of the concrete value of the several each factor to affect subitem comprised:

Overhaul management is because of several: overhaul type according in power grid maintenance scheme, and determine that overhaul management is because of several values, wherein, scheduled overhaul can not produce risk to electrical network, and its value is 1; And all there is possibility electrical network being brought to risk in non-plan repair and accident maintenance; That is, overhaul management because of the reference span of the several each factor to affect subitem comprised as follows:

Maintenance type (that is: overhaul management factor to affect subitem)	Overhaul management factor to affect subitem value
		Scheduled overhaul	1

Non-plan repair	1.5
		Accident maintenance	2

Repair time is because of several: determine that the repair time is because of several values according to repair rate in power grid maintenance scheme, wherein, repair rate is higher, then the probability of electrical network occurrence risk is lower, then the repair time because of several values less; That is, the repair time sees the following form because of the concrete value of reference of the several each factor to affect subitem comprised:

Repair rate (that is: repair time factor to affect subitem)	Repair time factor to affect subitem value
		1 ~ 3 day	0.3～0.5
3 ~ 10 days	0.5～1
		10 ~ 30 days	1～1.5
More than 30 days	1.5～3

Reference value or the reference span of each factor to affect subitem that factor to affect S comprises can be obtained thus.

S5, after the weighted value between each factor to affect subitem that the factor to affect item classification and often kind of factor to affect item that obtain participating in risk quantification assessment comprise; The electrical network hazard factor of quantitative evaluation as required, corresponding factor to affect subitem is chosen from the factor to affect item of each participation risk quantification assessment, each selected factor to affect subitem is multiplied and namely obtains probable value, again probable value is multiplied by consequence value, namely obtains power grid risk value when described hazard factor occurs electrical network.

Herein, the concrete factor to affect subitem participated in the factor to affect item of risk quantification assessment will be determined based on needing the electrical network hazard factor of quantitative evaluation, such as, after obtaining probable value=(device type factor * fault category factor * historical statistics factor * weather factor * equipment deficiency factor * overhaul management factor * repair time factor * site operation factor * control measure factor), for device type factor, if it comprises two sub-factors of impact altogether, be respectively: circuit weight is 0.47, unit weight is 0.54; If hazard factor is: need to overhaul power network line, then when calculating probability value, the value of device type factor is circuit weight, is 0.47; If hazard factor is: need to overhaul electrical network unit, then when calculating probability value, the value of device type factor is unit weight, is 0.54.

After obtaining value-at-risk, also comprise: the corresponding relation of predefine power grid risk value scope and power grid risk grade; By searching described corresponding relation, obtain the power grid risk grade corresponding to power grid risk value that S5 obtains; And show described power grid risk grade by patterned way.As a kind of reference, the relation of value-at-risk and risk class can be:

I grade of risk (jumbo line): 1000≤value-at-risk

II grade of risk (material risk): 300≤value-at-risk < 1000

III grade of risk (greater risk): 60≤value-at-risk < 300

IV grade of risk (general category-A risk): 20≤value-at-risk < 60

V grade of risk (general category-B risk): 5≤value-at-risk < 20

Above risk class can adjust according to actual conditions, adjustment according to need according to repeatedly to risk quantification assessment experiment and expertise adjust.

Adopting said method to calculate power grid risk grade to multiple hazard factor, when have accumulated a certain amount of sample data, the method for sorter can be adopted, simplify the deterministic process of risk class, that is:

S7, sets up preliminary classification device, using many groups corresponding relation of known hazard factor and power grid risk grade as training sample, uses described training sample to train described preliminary classification device, constantly the parameter of the described preliminary classification device of adjustment; Thus obtain the sorter that trains;

S8, choose different from described training sample many groups and verify sample, the input of the sorter using the hazard factor in described checking sample to obtain as S7, judge that whether the output of sorter is consistent with the power grid risk grade in checking sample, if consistent, then show that the sorter that S7 obtains passes through checking, perform S8; Otherwise, continue to choose other training samples, the sorter that S7 obtains trained further, until obtain the sorter by checking;

S9, when needs carry out risk class division to a certain hazard factor, directly described hazard factor is input to the described sorter by checking, described sorter directly exports final risk class.

Wherein, the sorting algorithm that sorter adopts is the combination of one or more in decision Tree algorithms, neural network algorithm, Bayes net algorithm, genetic algorithm and algorithm of support vector machine.

For the above-mentioned safe operation of electric network risk quantification method provided, can be occurred by the form of separate functional blocks, that is, a safe operation of electric network quantifying risk can be built and quantize platform, realize above-mentioned safe operation of electric network risk quantification method.Platform is quantized for this safe operation of electric network quantifying risk, is described below:

(1) technology path

Adopt the technical system meeting J2EE architecture standard to build, present application with B/S pattern.

Platform occurs with the form of a separate functional blocks, can the form of URL be articulated on DMIS, as the entrance of platform.Platform, by the Single Sign On of authority, can realize once logging in DMIS, gets final product the corresponding function in usage platform.

(2) network architecture

Adopt and concentrate deployment way (B/S).Data base manipulation has Oracle, opens up a user storage data, and middleware adopts the Tomcat that increases income.Server hardware utilizes existing hardware.Short message service can utilize total tune existing device and service.

Wherein, basic technology platform adopts J-HOP platform, and this platform adopts main flow J2EE technology to build, and forms performance application service based on middleware Technology, carries out background business logic process, foreground function demonstration and mutual.

Workflow platform that J-HOP is built-in, authority platform, form platform, regulation engine and log management platform, and externally provide infrastructure service in the form of services.

J-HOP provides the instruments such as workflow design, Report Form Design, Design with Rule, and by calling built-in service, for designing, developing, maintainer provides convenience, quick, patterned Software Engineering Management instrument.

(3) control of authority

single-sign-on

The same user of the locked resource of accessing in same server different application is realized by single-sign-on; only need to log in once; namely by after the safety verification in an application, when visiting again the locked resource in other application, login authentication again is no longer needed.

(4) version updating

Quantitative model supports that different editions coexists, and can realize the replacement training version, current version, old version.

From technological layer, platform carries out version management to quantitative model.The formal quantitative model of current issue, for running environment provides quantifying risk to quantize ability.And train Model of Version, then in running environment, carry out training adjustment and analysis verification, new official release becomes current version after issuing, and former current version becomes old version.

Version management is carried out to quantitative model, can follow the trail of the evolution process of model on the one hand, can standardize to the management of model on the other hand.

In Risk Management Platform, the version upgrading function of risk quantification model and mechanism, risk quantification can be realized from training version to the favourable conversions of current version, old version, different editions represents the regular period to the level of understanding of risk quantification work and requirement, thus makes platform have stronger adaptability and vitality.

Such as, for the factor to affect of computational scheme under blue heavy rain weather condition.

According to Bayesian formula, can obtain:

P(A|B)＝P(A∩B)/P(B)

Wherein:

A: line tripping event

B: blue heavy rain weather warning event

P (A): line tripping probability of occurrence

P (B): blue heavy rain weather warning probability of occurrence (this value can be similar to and adopt empirical value C0 to replace)

P (A|B): under blue heavy rain weather warning condition, line tripping probability of occurrence

P (A ∩ B): the probability that line tripping event occurs while blue weather warning

Remain with 2006-2010 in the supposing the system line fault tripping operation record N1 bar of totally 5 years, as:

Sequence number	Fault-time	Event of failure	Weather condition	…
					1	2005-09-10	* line trips	Fine	…
2	2005-09-15	* line trips	Blue heavy rain early warning	…
					…	…	…	…	…
N	2010-12-11	* line trips	Fine	…

The record wherein counting line tripping has M1 bar, and wherein blue heavy rain early warning weather condition line tripping operation records Q1 bar, and the tripping operation of normal weather situation line records R1.Note P ( a) be fault trip probability under normal circumstances, then:

P( A)＝R1/N1

P(A|B)＝P(A∩B)/P(B)＝(Q1/N1)/C0

In risk quantification model, final it is envisaged that for the normal weather situation line fault trip probability as fiducial value 1, the break down relative weighting Ki of tripping operation probability of blue heavy rain early warning situation line can be expressed as:

Ki＝P(A|B)/P( A)＝((Q1/N1)/C0)/(R1/N1)＝Q1/(R1*C0)

In the risk quantification model of current shape, although calculative coefficient comprises the relative weighting that fault trip probability occurs in normal weather situation at blue heavy rain early warning situation lower probability these equipment relative the various kinds of equipment such as circuit, bus, main transformer, genset.But because weather factor is for the impact of equipment, mainly for outdoor equipment, also just main is circuit, and therefore adopt the result of calculation of circuit to deduce here.

When needing to readjust this coefficient the end of the year 2011, the 2005-2011 tripping operation accumulated in Mishap Database records M2 bar, wherein blue heavy rain early warning weather condition line tripping operation records Q2 bar, and the tripping operation of normal weather situation line records R2, then new weight Ki ' can be expressed as:

Ki’＝Q2/(R2*C0)

Therefore, when considering new samples in 2011, the value of new weight is:

Ki’＝(Q2*R1)/(Q1*R2)*Ki

This is also on the whole based on the new coefficient updating method of Bayes's thought:

Old model+new samples=> new model

In reality, old COEFFICIENT K i may rule of thumb obtain, and may be to be obtained by southern net fault statistics data, also may be by obtaining from version more early in conjunction with sample.

(5) risk is issued

Platform have employed the various ways such as Risk statistic distribution plan, risk control Gantt chart, venture analysis radar map have carried out visual pattern expression and displaying to the total number of risk and distribution, risk management and control progress, venture influence factor size, be conducive to risk management related personnel to find and problem analysis, improve the service efficiency of platform.

In addition, in order to carry out scientific quantification to the risk size of various risk source, must analyze the possible hazardous events of risk source, estimate its harm size and probability of happening, a risk source may have multiple hazardous events, for this reason, and the extendible harm identification data model of Platform Designing, support, to the structure of the corresponding multiple hazardous events of a risk source, to meet the actual needs that harm identifies.

The size of risk depends on the size of each hazardous events, the process of risk quantification quantizes each may endanger fault harm consequence and fault rate, quantitative model is adopted to calculate during quantification, draw harm size and the probability of happening of each possibility hazardous events, and using the grade of the risk class of hazardous events the highest for risk rating as this risk, carry out risk prevention system work according to most serious conditions, avoid the generation of hazardous events to greatest extent.Risk Pre-control measure and requirement to be proposed simultaneously.

As can be seen here, safe operation of electric network risk quantification method provided by the invention, has the following advantages:

(1) when calculating each factor to affect subitem weight that often kind of factor to affect item comprises, be respectively electrical network year fault statistics table and third party's fault statistics table be provided with weight, not only take full advantage of the annual fault statistics table of electrical network and third party's fault statistics table, and, different expert can be met to electrical network year fault statistics table and the different degree of beliefs of third party's fault statistics table, realize the quantitative model version that different provinces and cities form self different characteristics.And whole quantum chemical method is succinctly quick, does not have the puzzlement of convergence problem, requires low to hardware environment, easily promote the use of;

(2) platform realize quantitative model version upgrading function and mechanism, subjectivity, the dynamic feature of social development requirement and people's risk understanding are taken into full account, support to adjust the quantitative criteria of risk schedule, the influence factor of risk probability is segmented, screens and accepted or rejected, quantitative model version can be realized from training version to the favourable conversions of current version, old version, different editions represents the regular period to the level of understanding of risk quantification work and requirement, thus makes platform have stronger adaptability and vitality;

(3) in power production process, method provided by the invention has science and the accuracy of higher risk quantification, reasonably can quantize risk, as the objective measure foundation instructing electric power actual production.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should look protection scope of the present invention.

Claims (8)

1. a safe operation of electric network risk quantification method, is characterized in that, comprises the following steps:

S1, sets up risk quantification model, and the risk quantification algorithm that described risk quantification model adopts, for connecting multiplication algorithm, obtains the power grid risk value universal calculation equation shown in formula one thus:

Power grid risk value=consequence value * probable value; Formula one

Wherein, consequence value is the value that the consequence brought after being occurred by power grid risk carries out quantizing, and its expression formula is: consequence value=consequence is seriously worth * social influence factor; Wherein, consequence be seriously worth refer to the loss load of evaluated electrical network to carry out quantification conversion after obtain numerical value; The value of social influence factor and the guarantor of evaluated electrical network power rank positive correlation, and it is higher to protect power supply rank, then the value of social influence factor is larger;

Probable value=factor to affect item A1* factor to affect item A2 ... factor to affect item Ab; B is the sum screening the factor to affect item obtained;

S2, the grid equipment fault database that described risk quantification model read is up-to-date, obtains some fault cases between a certain historical years T1 to current year T2, and each fault case records m1 bar factor to affect item when there is certain fault; Wherein, each factor to affect item includes several factor to affect subitems;

S3, determines the classification needing the factor to affect item participating in risk quantification assessment:

Be specially:

S3.1, described risk quantification model calculates the information entropy of every bar factor to affect item in m1 bar factor to affect item respectively;

S3.2, described risk quantification model sorts to m1 bar factor to affect item by information entropy order from high to low; Herein, information entropy reacts the significance level of corresponding factor to affect item; Then, the m2 bar factor to affect item be arranged in front is filtered out; Wherein, 1≤m2≤m1, makes the summation of the information entropy of filtered out m2 bar factor to affect item reach the number percent of set information amount; Described m2 bar factor to affect item is the factor to affect item of the participation risk quantification assessment filtered out;

S4, for filtered out each factor to affect item, it forms by several factor to affect subitems; Obtain the weighted value between each factor to affect subitem that each described factor to affect item comprises;

Wherein, for device type factor to affect item, the weighted value between the factor to affect subitem that it comprises obtains by the following method:

S4.1, supposes that described device type factor to affect item comprises z device type factor to affect subitem;

S4.2, reads the annual fault statistics table of electrical network and third party's fault statistics table; Wherein, the number of times of each device type factor to affect subitem generation forced outage corresponding to described electrical network annual fault statistics table record different year; The forced outage rate of each device type factor to affect subitem corresponding to described third party's fault statistics table record different year;

S4.3, obtain the record between historical years T1 to current year T2 in described electrical network year fault statistics table, then, calculate the average time that forced outage occurs often kind of device type factor to affect subitem between described T1 to T2 year, be designated as C1, C2 respectively ... Cz;

Obtain the record between historical years T1 to current year T2 in described third party's fault statistics table, then, calculate the mean value of often kind of device type factor to affect subitem forced outage rate between described T1 to T2 year, be designated as D1, D2 respectively ... Dz;

S4.4, if the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: fault mean value=(Ci*Q1+Di*Q2)/2 of often kind of device type factor to affect subitem;

S4.5, is normalized the fault mean value of calculated various device type factor to affect subitems, obtains the weighted value of various device type factor to affect subitem;

Described historical statistics factor to affect item is obtained by following formulae discovery:

Value=every the annual of 1+ same category of device of historical statistics factor to affect item is broken down the mean value of number of times/same category of device sum=1+ same category of device failure rate over the years;

S5, after the weighted value between each factor to affect subitem that the factor to affect item classification and often kind of factor to affect item that obtain participating in risk quantification assessment comprise; The electrical network hazard factor of quantitative evaluation as required, corresponding factor to affect subitem is chosen from the factor to affect item of each participation risk quantification assessment, each selected factor to affect subitem is multiplied and namely obtains probable value, again probable value is multiplied by consequence value, namely obtains power grid risk value when described hazard factor occurs electrical network;

S6, the corresponding relation of predefine power grid risk value scope and power grid risk grade; By searching described corresponding relation, obtain the power grid risk grade corresponding to power grid risk value that S5 obtains; And show described power grid risk grade by patterned way.

2. safe operation of electric network risk quantification method according to claim 1, it is characterized in that, consequence is seriously worth value by the following method:

The security incident grade of standard comprises 5 ranks, is respectively: especially big loss level, heavy losses rank, greater loss rank, general category-A loss level and general category-B loss level; The span that consequence corresponding to each rank is seriously worth sees the following form:

Rank Consequence is seriously worth value Especially big loss More than 3000 Heavy losses 500 < consequences are seriously worth≤and 3000 Greater loss 100 < consequences are seriously worth≤and 500 General category-A loss 30 < consequences are seriously worth≤and 100 General category-B loss 1 < consequence is seriously worth≤and 30

Obtain consequence thus to be seriously worth.

3. safe operation of electric network risk quantification method according to claim 1, it is characterized in that, social influence because of the several social influence factor subitem comprised for superfine protecting power supply factor to affect subitem, one-level protects power supply factor to affect subitem, secondary protects power supply factor to affect subitem, particular time protects power supply factor to affect subitem and general period factor to affect subitem; The value of each factor to affect subitem is:

Period Social influence factor value Superfine guarantor is powered 2 One-level protects power supply 1.6 Secondary protects power supply 1.4 Particular time protects power supply 1.2 General period 1

Obtain the value of each factor to affect subitem in social influence factor thus.

4. safe operation of electric network risk quantification method according to claim 1, it is characterized in that, S3 is specially:

Information entropy by following formulae discovery every bar factor to affect item:

$H_s=-\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{p}_i{\log }_2{p}_i$

Wherein, H _sfor the information entropy of factor to affect item, p _ifor the probability of malfunction of factor to affect subitem between a certain historical juncture T1 to current time T2; The sum of the factor to affect subitem that n comprises for factor to affect item.

5. safe operation of electric network risk quantification method according to claim 1, it is characterized in that, for the other types factor to affect item except device type factor to affect item and historical statistics factor to affect item participating in risk quantification assessment, referred to as factor to affect item S, then: for factor to affect item S, the weighted value between its each factor to affect subitem is obtained by following steps S4-6:

S4-6, in each factor to affect subitem that factor to affect item S comprises, chooses any one factor to affect subitem as baseline impact factor subitem, other influences factor subitem is called non-referenced factor to affect subitem;

Read the electrical network year fault statistics table between historical years T1 to current year T2, count baseline impact factor subitem occurrence number and total effective degree, draw baseline impact factor subitem probability of occurrence; Count non-referenced factor to affect subitem occurrence number and total effective degree again, draw non-referenced factor to affect subitem probability of occurrence; By each non-referenced factor to affect subitem probability of occurrence divided by baseline impact factor subitem probability of occurrence, obtain normalized non-referenced factor to affect subitem weight K1, wherein, baseline impact factor subitem weight is 1;

Read the third party's fault statistics table between historical years T1 to current year T2, calculate the year number of stoppages average value P 1 of often kind of non-referenced factor to affect subitem, and the year number of stoppages average value P 2, P1 of going back Calculation Basis factor to affect subitem obtains weight K2 after normalization divided by P2;

Again because the weight number percent of the annual fault statistics table of electrical network is Q1, the weight number percent of third party's fault statistics table is Q2; Wherein, Q1+Q2=1;

Then: weighted value=(K1*Q1+K2*Q2)/2 of often kind of non-referenced factor to affect subitem.

6. safe operation of electric network risk quantification method according to claim 5, is characterized in that, after calculating the weighted value in factor to affect item S between each factor to affect subitem, also comprises: the step verified calculated weighted value; If met with reference to span, or in the predetermined amplitude of reference value, then calculated weighted value is the weighted value finally calculated; Otherwise, expand the annual scope participating in the annual fault statistics table of electrical network and third party's fault statistics table adding up, and re-execute S4-6, until meet;

Concrete, factor to affect item S comprises site operation because several, control measure are because several, fault category is because several, weather effect is because several, equipment deficiency is because several, overhaul management is because several and repair time is because of the one or more combination in several;

Reference span or the reference value of the weighted value between each factor to affect subitem that factor to affect item S comprises are respectively:

Site operation comprises two factor to affect subitems because of several, is respectively: site environment does not use large-scale repair apparatus factor to affect subitem and site environment to use large-scale repair apparatus factor to affect subitem; Site environment does not use the reference value of large-scale repair apparatus factor to affect subitem to be 1; Site environment uses the span of large-scale repair apparatus factor to affect subitem to be: 1 < site environment uses span≤2 of large-scale repair apparatus factor to affect subitem;

Control measure comprise two factor to affect subitems because of several, are respectively: adopt spy's equipment of patrolling to carry out patrolling and examining and do not adopt spy's equipment of patrolling to patrol and examine; Wherein, when adopting the spy equipment of patrolling to patrol and examine, the probability of electrical network occurrence risk diminishes, then the value that the special equipment of patrolling carries out patrolling and examining is less than and does not adopt spy's equipment of patrolling to carry out the value of patrolling and examining; Further, 0< control measure are because of each factor to affect subitem span≤1 in several;

Fault category is because of several: bear the requirement determination fault category of the safety and stability standard of maximum perturbation ability according to power platform because of several values; Concrete, one-level, secondary and three level security stabilized references are corresponding first kind fault, Equations of The Second Kind fault and the 3rd class fault respectively; That is, fault category comprises three fault category factor to affect subitems because of several, and the value of each fault category factor to affect subitem sees the following form:

Fault category Fault category factor to affect subitem value First kind fault 1 Equations of The Second Kind fault 0.2～0.6 3rd class fault 0～0.2

Weather effect is because of several: according to the weather condition of maintenance time in power grid maintenance scheme, and determine that weather effect is because of several values, weather condition is more severe, and the probability of electrical network occurrence risk is higher, then weather effect because of several values larger; Weather effect sees the following form because of the concrete value of the several factor to affect subitem comprised:

Weather pattern (weather effect factor subitem) The value of weather effect factor subitem Normally 1 Thunderstorm gale 1～2 Risk of forest fire 1～1.5 High temperature 1～1.2 Dense fog 1～1.2 Heavy snow 1～4 Freeze 1～3

Equipment deficiency is because of several: according to the serious conditions of equipment maintained in power grid maintenance scheme whether existing defects and defect, and determine that equipment deficiency is because of several concrete values; Concrete, equipment deficiency is more serious, and the probability of generation power grid risk is higher, then equipment deficiency because of several values larger; Therefore, equipment deficiency sees the following form because of the concrete value of the several each factor to affect subitem comprised:

Overhaul management is because of several: overhaul type according in power grid maintenance scheme, and determine that overhaul management is because of several values, wherein, scheduled overhaul can not produce risk to electrical network, and its value is 1; And all there is possibility electrical network being brought to risk in non-plan repair and accident maintenance; That is, overhaul management because of the reference span of the several each factor to affect subitem comprised as follows:

Maintenance type (that is: overhaul management factor to affect subitem) Overhaul management factor to affect subitem value Scheduled overhaul 1 Non-plan repair 1.5 Accident maintenance 2

Repair time is because of several: determine that the repair time is because of several values according to repair rate in power grid maintenance scheme, wherein, repair rate is higher, then the probability of electrical network occurrence risk is lower, then the repair time because of several values less; That is, the repair time sees the following form because of the concrete value of reference of the several each factor to affect subitem comprised:

Repair rate (that is: repair time factor to affect subitem) Repair time factor to affect subitem value 1 ~ 3 day 0.3～0.5 3 ~ 10 days 0.5～1 10 ~ 30 days 1～1.5 More than 30 days 1.5～3

Reference value or the reference span of each factor to affect subitem that factor to affect S comprises can be obtained thus.

7. safe operation of electric network risk quantification method according to claim 1, is characterized in that, also comprise:

S7, sets up preliminary classification device, using many groups corresponding relation of known hazard factor and power grid risk grade as training sample, uses described training sample to train described preliminary classification device, constantly the parameter of the described preliminary classification device of adjustment; Thus obtain the sorter that trains;

S8, choose different from described training sample many groups and verify sample, the input of the sorter using the hazard factor in described checking sample to obtain as S7, judge that whether the output of sorter is consistent with the power grid risk grade in checking sample, if consistent, then show that the sorter that S7 obtains passes through checking, perform S8; Otherwise, continue to choose other training samples, the sorter that S7 obtains trained further, until obtain the sorter by checking;

S9, when needs carry out risk class division to a certain hazard factor, directly described hazard factor is input to the described sorter by checking, described sorter directly exports final risk class.

8. safe operation of electric network risk quantification method according to claim 7, it is characterized in that, the sorting algorithm that described sorter adopts is the combination of one or more in decision Tree algorithms, neural network algorithm, Bayes net algorithm, genetic algorithm and algorithm of support vector machine.