CN102945314A - Method for computing constant failure rate of self-adapting relay protection device - Google Patents

Abstract

The invention provides a method for computing constant failure rate of a self-adapting relay protection device, which belongs to the technical field of electric system protection. The method comprises the following steps: acquiring measured data on site; computing the total number of invalid devices, and if the total number of the invalid devices is less than 10, judging the failure rate which is shown in the specification, in which n is the sum of the invalid times and Ti is the operation time of a numbered i failing device since the last failure; if the total number of the invalid devices is larger than or equal to 10, continuing to judge the type of the measured data on site, if the invalid data is complete data, determining no termination, and judging the failure rate which is shown in the specification, in which Nst is the total number of protection devices at the monument t and Ns delta t is the number of invalid protection devices after the monument delta t; if censored data is contained in the measured invalid data on site, judging the type of the censored data, and if left censored data is existed, judging the failure rate which is shown in the specification; and if left censored data is not existed, performing termination at fixed time, and judging the failure rate which is shown in the specification. By adopting the method provided by the invention, failure rate estimation of various measured data on site can be realized, and the computation of the failure rate is more real and reliable.

Description

A kind of adaptive protective relaying device constant failure-rate computing method

Technical field

The invention belongs to the protecting electrical power system technical field, relate in particular to adaptive protective relaying device constant failure-rate (crash rate) computing method based on measured data.

Background technology

Protective relaying device is the visual plant that guarantees electric power system safe operation and reliable power supply.When electrical network breaks down; if protective relaying device can not correct operation; the fault of electric system is enlarged; even may cause because of bad chain reaction whole mains breakdown; cause large-area power-cuts, bring to people's same normal life, economic development and social stability to have a strong impact on.Therefore, the reliability of guaranteeing protective relaying device is a very important content that ensures power network safety operation.Relay protection system is carried out Reliability Analysis Research, can in time find the hidden danger of protective relaying device, help to eliminate fast the protection defective, improve the available rate of protection, cause the expansion of accident when avoiding electrical network to have an accident because of the protective device incorrect operation.

The definition of protective relaying device when operation reliability index and calculating are calculated with For The Reliability Indicas of Gereration System, evaluation, the use of protective relaying device, are improved and develop etc. closely related in the protecting electrical power system.By the reliability of relay protection assessment, get the various reliability indexs such as failure rate, repair rate and availability, design model, according to model, draw the relay protection device optimal repair cycle, to instruct power department reasonable arrangement turnaround plan.

At present, the reliability model of protective relaying device roughly is divided into two classes, and the one, adopt Monte Carlo Analogue Method, take law of great numbers as foundation, to the analog computation of relay protection device available rate; The another kind of analytical method that is referred to as, as take fault tree as the basis fault tree analysis and the Markov model take Markov process as foundation.The estimation of protective relaying device failure rate also is the prerequisite and basis of many protective relaying device reliability estimation methods such as Markov model.

But; for the field measurement fail data; the fail data record of protective relaying device is varied, and many, the Time Inconsistency that puts into operation of protective device quantity of on-the-spot operation, and the Censored Data problem of left truncation and Censoring can appear in some measured data.Therefore, for different fail datas and truncation situation, should there be a kind of adaptive method to calculate the failure rate of the protective relaying device under the different situations.But; existing article about protective relaying device crash rate evaluation method seldom; not yet propose a kind of adaptive crash rate evaluation method based on measured data, therefore, on-the-spot measured data still can not be well by the estimation of program realization to the protective relaying device failure rate.

Summary of the invention

The present invention is directed in the method for existing protective relaying device failure rate estimation, lack a kind of can be compatible and be applied to the method for multiple field measurement data, a kind of adaptive protective relaying device constant failure-rate computing method based on measured data have been proposed.The method can be applied to the failure rate estimation of various field measured data, so that the calculating of failure rate is more true and reliable.

Technical scheme is, a kind of adaptive protective relaying device constant failure-rate computing method, and described method comprises the following steps:

Step 1: obtain field measurement data (fault start time, fault apart from fault-time last time, fault concluding time, device sum and the disabling device number etc. that comprise protective relaying device);

Step 2: process measured data, calculate fail data number of times and fault apart from information such as fault-times last time;

Step 3: the censored data type of judging the field measurement data;

Step 4: the failure rate λ computing method different with the fail data selection of times according to different censored data types.

In the censored data type of described judgement field measurement data, comprise partial data, left censored data and three kinds of censored data types of Type I censoring data.

Described calculating fail data number of times is used for judging which kind of evaluation method is the field measurement data can be applicable to, such as the data approximating method.

Described different failure rate λ computing method comprise method in following 4:

(1) Failure count is less than 10 o'clock, employing method 1;

(2) Failure count is greater than 10 o'clock, and do not have censored data, employing method 2;

(3) there was left censored data in Failure count greater than 10 o'clock, employing method 3;

(4) Failure count is greater than 10, and only has Type I censoring data, employing method 4.

In the described different failure rate λ computing method, method 1 is for asking for the method for inverse working time before the average inefficacy. , wherein n is the fail data summation, T _iBe that the device of i platform fault is apart from the time of failure operation last time.

In the described different failure rate λ computing method, method 2 is for considering the Maximum Likelihood Estimation Method of repair time and Type I censoring data. , N wherein _StProtective device is total constantly for t, N _{S Δ t}Protective device number for Δ t inefficacy after the moment.

In the described different failure rate λ computing method, method 3 is the method for not considering the least square fitting of truncation.

, t wherein _iFor losing efficacy constantly F (t _i) be the cumulative distribution function of protective relaying device.

In the described different failure rate λ computing method, the method that method 4 is estimated for the maximum likelihood of considering left truncation.

, wherein n is the Failure count summation, n _CenBe censored data summation, T _iBe that the device of i platform fault is apart from the time of failure operation last time.

The present invention can be applied to the failure rate estimation of various field measured data, and the calculating of failure rate is more true and reliable.

Description of drawings

Fig. 1 is the adaptive approach process flow diagram of protective relaying device;

Embodiment

Below in conjunction with accompanying drawing, preferred embodiment is elaborated.Should be emphasized that following explanation only is exemplary, rather than in order to limit the scope of the invention and to use.

Fig. 1 is the adaptive approach process flow diagram of protective relaying device.As shown in Figure 1, a kind of adaptive protective relaying device constant failure-rate computing method, the method step is as follows:

Step 1: obtain field measurement data (fault start time, fault apart from fault-time last time, fault concluding time, device sum and the disabling device number etc. that comprise protective relaying device);

It is poor that fault concluding time of protective relaying device and start time are done, and can obtain the repair time; Fault start time and fault were done the time that difference can get time operation of the upper fault of auto levelizer distance apart from fault-time last time.

Step 2: process measured data, calculate fail data number of times and fault apart from information such as fault-times last time;

The fail data number of times is the disabling device sum, if the fail data sum of statistics very little, then can't finish the condition of maximum likelihood estimation and data least square fitting, here add up the fail data number of times, purpose is to seek a threshold value, guarantee the result of result that estimated value is estimated and data fitting more near actual value, and short fail data number of times can be estimated or the Fitting Calculation.

Step 3: the censored data type of judging the field measurement data;

If the data number of times, namely the device number of inefficacy equates with the device sum, illustrates that this time statistics is for adding up fully, and there is not the truncation situation in data; If fail data ends in a certain date, then explanation is this time added up and is had Censoring, and much all there is Censoring in general field data, because all protective devices all can not be come out fault-time; If some protective device puts into operation too early, then may lack fail data, just thereby know that there is the impact of left truncation in this statistics.If there is truncation in data and does not consider the impact that truncation brings, then the error of calculation of failure rate can be larger.

Step 4: the failure rate computing method different with the fail data selection of times according to different censored data types.

When the fail data number of times was less, the method that adopts match or maximum likelihood to estimate had limitation, this moment can only by failure rate λ and Mean Time Between Failures each other relation reciprocal calculate, so can adopt method 1, namely

${\mathrm{\λ}}_1=\frac{1}{\frac{1}{n}{\displaystyle \underset{i=1}{\overset{n}{\∑}}{T}_i}}---\left(1\right)$

When the fail data number of times is a lot, just need not to consider above-mentioned impact, more accurate for making calculating, utilize more fully the field measurement data, the calculating of failure rate is divided into three kinds of partial data, left censored data and Type I censoring datas.In the evaluation method of failure rate, least square method is not subjected to the impact of Censoring, and Maximum Likelihood Estimation Method is not subjected to the impact of left truncation.

When fail data was partial data, the method that can adopt maximum likelihood to estimate realized that namely method 2.The result that the method is estimated is the definition of failure rate λ: in a period of time, and the probability that protective relaying device breaks down, so formula is

${\mathrm{\&lambda;}}_2=\frac{P\{t<T<t+\mathrm{\&Delta;}t\}}{\mathrm{\&Delta;}t}=\frac{N_{s\mathrm{\&Delta;}t}}{\mathrm{\&Delta;}t{N}_{st}}---\left(2\right)$

Generally speaking, data all have Censoring, when fail data contains left truncation, can adopt the method for least square fitting to realize, namely method 3.Satisfy exponential distribution because protective relaying device is approximate, can establish F (t _i) be the cumulative distribution function of protective relaying device, then

$F(t_i)={\displaystyle {\&Integral;}_{t_i}^{\mathrm{\&infin;}}\mathrm{\&lambda;}{e}^{\mathrm{\&lambda;}t}dt}=1?e^{\mathrm{\&lambda;}{t}_i}---\left(3\right)$

Formula (3) is put in order and take the logarithm in both sides, can get

$\ln (1?F(t_i))=\mathrm{\&lambda;}{t}_i---\left(4\right)$

Utilize the formula of linear function least square fitting, can obtain the computing formula of failure rate λ

${\mathrm{\&lambda;}}_3=\frac{{\displaystyle \underset{i=1}{\overset{n}{\&Sum;}}?t_i\ln (1?F(t_i))}}{{\displaystyle \underset{i=1}{\overset{n}{\&Sum;}}{t}_i^2}}---\left(5\right)$

When fail data only is Censoring, can adopt the method for the maximum likelihood estimation of considering Type I censoring data, namely method 4.The formula form when just calculating the mean free error time, is not calculated the Censoring number suc as formula (1) at this moment

${\mathrm{\&lambda;}}_4=\frac{n?n_{cen}}{{\displaystyle \underset{i=1}{\overset{n}{\&Sum;}}{T}_i}}---\left(6\right)$

Specific embodiment 1:

Certain electrical network model protective relaying device fail data is as follows:

The protective relaying device sum: 231, left censored data 42, fail data 188, Type I censoring data is 231-188=43, partial data is 231-42-43=146.Concrete protective relaying device out-of-service time and the time of putting into operation are seen the fail data form.

Employing is asked for the method for mean free error time inverse and is determined protective relaying device failure rate λ (method 1).

When relay protection fail data less (only be in form 43-51 capable 9 data such as fail data), then adopt method 1,

Specific embodiment 2:

Employing is asked for the method for mean free error time inverse and is determined protective relaying device failure rate λ (method 2).

Abundant when the relay protection fail data, and for partial data (be in form data 43 row after such as fail data), then adopt method 2,

Specific embodiment 3:

Employing is asked for the method for mean free error time inverse and is determined protective relaying device failure rate λ (method 3).

Abundant when the relay protection fail data, and there is left censored data (be total data in form such as fail data), then adopt method 3,

Specific embodiment 4:

Employing is asked for the method for mean free error time inverse and is determined protective relaying device failure rate λ (method 4).

When the relay protection fail data abundant, and do not have left censored data and exist Type I censoring data (as fail data be in the form data after 43 row and consider in the timing statistics not 43 equipment of fault), then adopt method 4,

The present invention can be applied to the failure rate estimation of various field measured data, and the calculating of failure rate is more true and reliable.

The above; only for the better embodiment of the present invention, but protection scope of the present invention is not limited to this, anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.

The fail data form

Claims (1)

1. adaptive protective relaying device constant failure-rate computing method is characterized in that described method comprises the following steps:

Step 1: obtain the field measurement data

Described data comprise fault start time, fault apart from fault-time last time, fault concluding time, device sum and the disabling device number of protective relaying device;

In the described field measurement data, data type can be various ways, both can be the Failure count of protective relaying device, also can be the out-of-service time of protective relaying device;

Described field measurement data type is divided into partial data, left censored data and Type I censoring data three classes with the field measurement data;

Step 2: calculate total disabling device number, if total disabling device is counted less than 10, failure rate λ then:

, wherein n is the Failure count summation, T _iBe that the device of i platform fault is apart from the time of failure operation last time;

Step 3: if total disabling device number, continues judgement field measurement data type more than or equal to 10, if fail data is partial data, namely do not have ending, then failure rate λ:

, N wherein _StProtective device is total constantly for t, N _{S Δ t}Protective device number for Δ t inefficacy after the moment;

Step 4: if the field measurement fail data exists censored data then to distinguish the censored data type, if there is left censored data, failure rate λ then:

, t wherein _iFor losing efficacy constantly F (t _i) be the cumulative distribution function of protective relaying device;

If there is no left censored data, i.e. Censoring, then failure rate λ:

, wherein n is the Failure count summation, n _CenBe censored data summation, T _iBe that the device of i platform fault is apart from the time of failure operation last time.