CN103995921A - Method for simulating and assessing micro-grid power supply system reliability - Google Patents

Abstract

The invention discloses a method for simulating and assessing micro-grid power supply system reliability. The method includes the three steps that (S1) an assessment model is determined, (S2) element state sampling is carried out, and (S3) system assessment is carried out. Historical operation data of various kinds of equipment serve as bases, distributed power supply fault points and fault time sequences are led in for analysis of system power supply reliability, the operation state of system distributed power supply equipment is simulated, influences of performance faults of the equipment on system power supply reliability are fully taken into consideration, and the analysis result can reflect power supply reliability of a micro-grid system truly.

Description

A kind of simulation evaluation method of microgrid electric power system reliability

Technical field

The present invention relates to a kind of simulation evaluation method of microgrid electric power system reliability.

Technical background

Microgrid power supply technique develops rapidly at present, but less to the research by microgrid electric power system reliability aspect.In the ripe software of existing microgrid electric power system emulation, generally to use the concept of load satisfaction to weigh the reliability that micro-grid system is powered, it is mainly by the request for utilization of the configuration of microgrid and microgrid engineering, to the satisfaction of loading of the ratio calculation micro-grid system of section system power supply total amount and system load demand total amount sometime, it has the following disadvantages: only with meeting load (Unmet electric load), do not weigh the reliability that micro-grid system is powered, as homer software, this software has only been considered the deficiency of load power supply, do not relate to the impact of burning natural gas distributed power apparatus fault on system power supply reliability, its result is accurate not.

Summary of the invention

Technical matters to be solved by this invention, just be to provide a kind of simulation evaluation method of microgrid electric power system reliability, the method is usingd the history data of all kinds of burning natural gas distributed power apparatus of microgrid electric power system as basis, running status to each burning natural gas distributed power apparatus is simulated, taken into full account each impact of burning natural gas distributed power apparatus faults itself on microgrid electric power system power supply reliability, the result obtaining is more accurate.

Solve the problems of the technologies described above, the technical solution used in the present invention is:

A simulation evaluation method for microgrid electric power system reliability, comprises the following steps:

S1 determines assessment models;

The sampling of S2 element state;

S3 carries out system evaluation;

Described step S1 comprises following sub-step:

S1-1 sets up blower fan model

$P_w=\left\{\begin{array}{cc}0,& 0\&le;{\mathrm{SW}}_t<V_{\mathrm{ci}}\\ (A+B\&times;{\mathrm{SW}}_t+C\&times;{\mathrm{SW}}_t^2)P_r,& V_{\mathrm{ci}}\&le;{\mathrm{SW}}_t<V_r\\ P_r,& V_r\&le;{\mathrm{SW}}_t\&le;V_{\mathrm{co}}\\ 0,& {\mathrm{SW}}_t>V_{\mathrm{co}}\end{array}\right.;---\left(1\right)$

In formula, P _wfor exerting oneself in real time of blower fan, unit is kW, and parameter A, B, C are the coefficient of polynomial fitting of blower fan power curve non-linear partial, SW _tbe the real-time wind speed of t hour, unit is m/s, V _cifor starting wind speed, V _rfor wind rating, V _cofor excision wind speed;

SW wherein _tthe sequence of exerting oneself in real time adopt autoregressive moving average (ARMA) model to produce:

SW _t＝μ _t+σ _ty _t； (2)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…-α _t-mθ _m； (3)

μ _tfor the historical mean wind speed in somewhere, σ _tfor the standard deviation of wind speed profile, y _tfor time series, φ _i(i=1 ... n) be autoregressive coefficient, θ _j(j=1 ... m) be running mean coefficient, α _tfor white noise coefficient, obey average and be 0, variance is independent normal distribution;

By formula (1), (2), (3), obtain horal air speed value in the whole year, and then try to achieve in the whole year the horal blower fan sequence of exerting oneself in real time;

S1-2 sets up photovoltage model

$P_b=\left\{\begin{array}{cc}{P}_{\mathrm{sn}}(G_{\mathrm{bt}}^2/\left(G_{\mathrm{std}}{R}_c\right)),& 0\&le;G_{\mathrm{bt}}<R_c\\ P_{\mathrm{sn}}(G_{\mathrm{bt}}/G_{\mathrm{std}}),& R_c\&le;G_{\mathrm{bt}}<G_{\mathrm{std}}\\ P_{\mathrm{sn}},& G_{\mathrm{bt}}\&GreaterEqual;G_{\mathrm{stf}}\end{array}\right.;---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit is kW; P _snfor the rated power of photovoltaic, be illustrated in the power that under standard test condition, unit light intensity can produce; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor a certain particular light intensity, under this intensity of illumination, photovoltaic is exerted oneself and is started from the non-linear linearity that becomes with the relation of intensity of illumination; G _btbe the real-time lighting intensity of t hour, unit is kW/m ²;

G _btreal-time lighting intensity by the sampling of the probability distribution statistical of historical light intensity is obtained;

S1-3 sets up battery model

$B_{t+1}=\left\{\begin{array}{cc}{B}_{\min },& B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\min }\\ B_t+{\mathrm{\&Delta;W}}_t& B_{\min }<B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\max }\\ B_{\max },& B_{\max }<B_t+{\mathrm{\&Delta;W}}_t\end{array}\right.;---\left(5\right)$

In formula, Δ W _tfor t outside charge/discharge electricity amount of accumulator in the period, it equals to discharge and recharge the product of power and period t; B _tfor discharging and recharging the residual capacity of front accumulator, B _t=B _norm* Soc (t), wherein B _normfor battery rating, Soc (t) is the state-of-charge before discharging and recharging, B _t+1for discharging and recharging the residual capacity that finishes rear accumulator; B _min, B _maxbe respectively maximum, the minimum capacity of accumulator;

S1-4 sets up load model

L _t＝L _p×P _w×P _d×P _h(t)；

In formula, L _pfor year load peak, P _wfor the value in year-all load curves corresponding with t hour, P _dfor the value in the week-daily load curve corresponding with t hour, P _h(t) be the value in the day-hour load curve corresponding with t hour; Load value when Lt is t;

Described step S2 comprises following sub-step:

S2-1 asks the front average operating time of component failure and mean repair time

In micro-grid system, most elements are repairable elements, and the cyclic process of " operation-stop transport-operation " that its state variation situation can be by stable state is simulated, and the parameter of element meets following relational expression:

$\left\{\begin{array}{c}\mathrm{MTTR}=1/\mathrm{\&lambda;}\\ \mathrm{MTTF}=1/\mathrm{\&mu;}\end{array}\right.;$

In formula, MTTF is average operating time before component failure, the mean repair time that MTTR is element, is respectively the average of TTF and TTR; λ is crash rate, i.e. in Failure count/year, μ is repair rate, repairs number of times/year, and TTF is working time before losing efficacy, i.e. time between failures, and TTR is repair time, i.e. the out-of-service time;

S2-2 asks the time of front working time of the inefficacy of element and reparation

When micro-grid system is carried out to fail-safe analysis, it has been generally acknowledged that the next fault of element and fault are last time irrelevant, element fault has without memory; Therefore, think the equal obeys index distribution of time of working time and reparation before the inefficacy of element, its probability density function is:

$\left\{\begin{array}{c}f\left(t\right)={\mathrm{\&lambda;e}}^{-\mathrm{\&lambda;t}}\\ g\left(t\right)={\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, f (t) represents the probability that element breaks down constantly at t; G (t) represents the probability that element has been repaired constantly at t; F (t) with the probability distribution function of g (t) is:

$\left\{\begin{array}{c}F\left(t\right)={1-e}^{-\mathrm{\&lambda;t}}\\ G\left(t\right)={1-\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, F (t) represents that element fault moment is less than the probability of t; G (t) represents that element reparation is less than the probability of t constantly;

Above formula is slightly done to change:

$\left\{\begin{array}{c}F\text{'}\left(t\right)={1-F=e}^{-\mathrm{\&lambda;t}}\\ G\text{'}\left(t\right)={1-G=\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.$

Wherein, F ' (t) represents the probability that the time between failures of element is t; G ' (t) represents the probability that repair time of element is t;

So, F ' (t) and G ' (t) be the number in interval [0,1]; Therefore, be positioned at the mode of the random number between [0,1] by generation, conversely the time between failures of sampling element and repair time, its sampling formula is:

$\left\{\begin{array}{c}\mathrm{TTF}=-\frac{1}{\mathrm{\&lambda;}}1n{R}_1\\ \mathrm{TTR}=-\frac{1}{\mathrm{\&mu;}}1n{R}_2\end{array}\right.;$

In formula, R ₁, R ₂be equally distributed random number between [0,1]; Adopt said method, to TTF and TTR difference alternate sampling;

Described step S3 comprises following sub-step:

S3-1 initialization simulated clock simulation clock is 0, produces at random m the random number between 0-1, according to the crash rate parameter lambda in each element state model try to achieve m non-failure operation time TTF, TTF _ithe TTF that represents i element;

S3-2 finds out minimum TTF _i;

S3-3 produces a random number to i element, and according to its repair rate parameter, μ tries to achieve fault correction time TTR _i;

S3-4 reads FMEA table, the load point affecting while searching element i fault, record these dead electricity load point frequency of power cut, power off time, lack the information such as delivery.

S3-5 produces a new random number, is translated into the running time T TF that element i is new _i';

S3-6 judges that whether simulated clock simulation clock is across year, not across the load point power failure information of record being added in year then in load point reliability index; If across year, adopt across year formula, calculate load point reliability index and the Reliability Index of this year;

Judging whether simulated clock simulation clock has been advanced to meets the required time span of Evaluation accuracy, if reach, performs step S3-7, does not reach and returns to step S3-2;

S3-7 simulation process finishes, and adds up the load point in each simulation year and the reliability index of system;

The reliability index average of S3-8 and then calculating whole system.

Beneficial effect: the present invention uses the history data of various kinds of equipment as basis, introduce distributed power source trouble spot, fault-time sequence and carry out the analysis of system power supply reliability, running status to system burning natural gas distributed power apparatus is simulated, taken into full account the impact of breaking down on system power supply reliability because of equipment self performance, its analysis result can reflect the power supply reliability of micro-grid system more really.

Accompanying drawing explanation

The outage model of Fig. 1 repairable elements;

The state variation cyclic process figure of Fig. 2 repairable elements;

The process flow diagram of Fig. 3 microgrid electric power system reliability assessment.

Embodiment

Referring to Fig. 1 to Fig. 3.

The simulation evaluation method embodiment of microgrid electric power system reliability of the present invention, comprises the following steps:

S1 determines assessment models;

The sampling of S2 element state;

S3 carries out system evaluation;

Described step S1 comprises following sub-step:

S1-1 sets up blower fan model

$P_w=\left\{\begin{array}{cc}0,& 0\&le;{\mathrm{SW}}_t<V_{\mathrm{ci}}\\ (A+B\&times;{\mathrm{SW}}_t+C\&times;{\mathrm{SW}}_t^2)P_r,& V_{\mathrm{ci}}\&le;{\mathrm{SW}}_t<V_r\\ P_r,& V_r\&le;{\mathrm{SW}}_t\&le;V_{\mathrm{co}}\\ 0,& {\mathrm{SW}}_t>V_{\mathrm{co}}\end{array}\right.;---\left(1\right)$

In formula, P _wfor exerting oneself in real time of blower fan, unit is kW, and parameter A, B, C are the coefficient of polynomial fitting of blower fan power curve non-linear partial, SW _tbe the real-time wind speed of t hour, unit is m/s, V _cifor starting wind speed, V _rfor wind rating, V _cofor excision wind speed;

SW wherein _tthe sequence of exerting oneself in real time adopt autoregressive moving average (ARMA) model to produce:

SW _t＝μ _t+σ _ty _t； (2)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…-α _t-mθ _m； (3)

μ _tfor the historical mean wind speed in somewhere, σ _tfor the standard deviation of wind speed profile, y _tfor time series, φ _i(i=1 ... n) be autoregressive coefficient, θ _j(j=1 ... m) be running mean coefficient, α _tfor white noise coefficient, obey average and be 0, variance is independent normal distribution;

These parameters can obtain by this area's historical wind speed statistics above; By formula (1), (2), (3), can obtain horal air speed value in the whole year, and then try to achieve in the whole year the horal blower fan sequence of exerting oneself in real time;

S1-2 sets up photovoltage model

$P_b=\left\{\begin{array}{cc}{P}_{\mathrm{sn}}(G_{\mathrm{bt}}^2/\left(G_{\mathrm{std}}{R}_c\right)),& 0\&le;G_{\mathrm{bt}}<R_c\\ P_{\mathrm{sn}}(G_{\mathrm{bt}}/G_{\mathrm{std}}),& R_c\&le;G_{\mathrm{bt}}<G_{\mathrm{std}}\\ P_{\mathrm{sn}},& G_{\mathrm{bt}}\&GreaterEqual;G_{\mathrm{stf}}\end{array}\right.;---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit is kW; P _snfor the rated power of photovoltaic, be illustrated in the power that under standard test condition, unit light intensity can produce; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor a certain particular light intensity, under this intensity of illumination, photovoltaic is exerted oneself and is started from the non-linear linearity that becomes with the relation of intensity of illumination; G _btbe the real-time lighting intensity of t hour, unit is kW/m ²;

G _btreal-time lighting intensity can be by the sampling of the probability distribution statistical of historical light intensity be obtained;

S1-3 sets up battery model

$B_{t+1}=\left\{\begin{array}{cc}{B}_{\min },& B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\min }\\ B_t+{\mathrm{\&Delta;W}}_t& B_{\min }<B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\max }\\ B_{\max },& B_{\max }<B_t+{\mathrm{\&Delta;W}}_t\end{array}\right.;---\left(5\right)$

In formula, Δ W _tfor t outside charge/discharge electricity amount (discharging and recharging the product of power and period t) of accumulator in the period; B _tfor discharging and recharging the residual capacity of front accumulator, B _t=B _norm* Soc (t), wherein B _normfor battery rating, Soc (t) is the state-of-charge before discharging and recharging, B _t+1for discharging and recharging the residual capacity that finishes rear accumulator; B _min, B _maxbe respectively maximum, the minimum capacity of accumulator;

S1-4 sets up load model

L _t＝L _p×P _w×P _d×P _h(t)；

In formula, L _pfor year load peak, P _wfor the value in year-all load curves corresponding with t hour, P _dfor the value in the week-daily load curve corresponding with t hour, P _h(t) be the value in the day-hour load curve corresponding with t hour; Load value when Lt is t.

Described step S2 comprises following sub-step:

S2-1 asks the front average operating time of component failure and mean repair time

In micro-grid system, most elements are repairable elements, and the cyclic process of " operation-stop transport-operation " that its state variation situation can be by stable state is simulated, as shown in Figure 2; These parameters meet following relational expression:

$\left\{\begin{array}{c}\mathrm{MTTR}=1/\mathrm{\&lambda;}\\ \mathrm{MTTF}=1/\mathrm{\&mu;}\end{array}\right.;$

In formula, MTTF is average operating time before component failure, the mean repair time that MTTR is element, is respectively the average of TTF and TTR; λ is crash rate (Failure count/year), and μ is repair rate (repairing number of times/year), and TTF is working time (time between failures) before losing efficacy, and TTR is repair time (out-of-service time);

S2-2 asks the time of front working time of the inefficacy of element and reparation

When micro-grid system is carried out to fail-safe analysis, it has been generally acknowledged that the next fault of element and fault are last time irrelevant, element fault has without memory; Therefore, can think the equal obeys index distribution of time of working time and reparation before the inefficacy of element, its probability density function is:

$\left\{\begin{array}{c}f\left(t\right)={\mathrm{\&lambda;e}}^{-\mathrm{\&lambda;t}}\\ g\left(t\right)={\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, f (t) represents the probability that element breaks down constantly at t; G (t) represents the probability that element has been repaired constantly at t; F (t) with the probability distribution function of g (t) is:

$\left\{\begin{array}{c}F\left(t\right)={1-e}^{-\mathrm{\&lambda;t}}\\ G\left(t\right)={1-\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, F (t) represents that element fault moment is less than the probability of t; G (t) represents that element reparation is less than the probability of t constantly;

Above formula is slightly done to change, can obtain:

$\left\{\begin{array}{c}F\text{'}\left(t\right)={1-F=e}^{-\mathrm{\&lambda;t}}\\ G\text{'}\left(t\right)={1-G=\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.$

Wherein, F ' (t) represents the probability that the time between failures of element is t; G ' (t) represents the probability that repair time of element is t;

Can find out, F ' (t) and G ' (t) be the number in interval [0,1]; Therefore, can be positioned at by generation the mode of the random number between [0,1], conversely the time between failures of sampling element and repair time, its sampling formula is:

$\left\{\begin{array}{c}\mathrm{TTF}=-\frac{1}{\mathrm{\&lambda;}}1n{R}_1\\ \mathrm{TTR}=-\frac{1}{\mathrm{\&mu;}}1n{R}_2\end{array}\right.;$

In formula, R ₁, R ₂be equally distributed random number between [0,1]; Adopt said method, to TTF and TTR difference alternate sampling;

Described step S3 comprises following sub-step:

S3-1 initialization simulated clock simulation clock is 0, produces at random m the random number between 0-1, according to the crash rate parameter lambda in each element state model try to achieve m non-failure operation time TTF, TTF _ithe TTF that represents i element;

S3-2 finds out minimum TTF _i;

S3-3 produces a random number to i element, and according to its repair rate parameter, μ tries to achieve fault correction time TTR _i;

S3-4 reads FMEA table, the load point affecting while searching element i fault, record these dead electricity load point frequency of power cut, power off time, lack the information such as delivery.

S3-5 produces a new random number, is translated into the running time T TFi' that element i is new;

S3-6 judges that whether simulated clock simulation clock is across year, not across the load point power failure information of record being added in year then in load point reliability index; If across year, adopt across year formula, calculate load point reliability index and the Reliability Index of this year;

Judging whether simulated clock simulation clock has been advanced to meets the required time span of Evaluation accuracy, if reach, performs step S3-7, does not reach and returns to step S3-2;

S3-7 simulation process finishes, and adds up the load point in each simulation year and the reliability index of system;

The reliability index average of S3-8 and then calculating whole system.

Claims (3)

1. a simulation evaluation method for microgrid electric power system reliability, is characterized in that comprising the following steps:

S1 determines assessment models;

The sampling of S2 element state;

S3 carries out system evaluation;

Described step S1 comprises following sub-step:

S1-1 sets up blower fan model

$P_w=\left\{\begin{array}{cc}0,& 0\&le;{\mathrm{SW}}_t<V_{\mathrm{ci}}\\ (A+B\&times;{\mathrm{SW}}_t+C\&times;{\mathrm{SW}}_t^2)P_r,& V_{\mathrm{ci}}\&le;{\mathrm{SW}}_t<V_r\\ P_r,& V_r\&le;{\mathrm{SW}}_t\&le;V_{\mathrm{co}}\\ 0,& {\mathrm{SW}}_t>V_{\mathrm{co}}\end{array}\right.;---\left(1\right)$

In formula, P _wfor exerting oneself in real time of blower fan, unit is kW, and parameter A, B, C are the coefficient of polynomial fitting of blower fan power curve non-linear partial, SW _tbe the real-time wind speed of t hour, unit is m/s, V _cifor starting wind speed, V _rfor wind rating, V _cofor excision wind speed;

SW wherein _tthe sequence of exerting oneself in real time adopt autoregressive moving-average model to produce:

SW _t＝μ _t+σ _ty _t； (2)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…-α _t-mθ _m； (3)

μ _tfor the historical mean wind speed in somewhere, σ _tfor the standard deviation of wind speed profile, y _tfor time series, φ _i(i=1 ... n) be autoregressive coefficient, θ _j(j=1 ... m) be running mean coefficient, α _tfor white noise coefficient, obey average and be 0, variance is independent normal distribution;

By formula (1), (2), (3), obtain horal air speed value in the whole year, and then try to achieve in the whole year the horal blower fan sequence of exerting oneself in real time;

S1-2 sets up photovoltage model

$P_b=\left\{\begin{array}{cc}{P}_{\mathrm{sn}}(G_{\mathrm{bt}}^2/\left(G_{\mathrm{std}}{R}_c\right)),& 0\&le;G_{\mathrm{bt}}<R_c\\ P_{\mathrm{sn}}(G_{\mathrm{bt}}/G_{\mathrm{std}}),& R_c\&le;G_{\mathrm{bt}}<G_{\mathrm{std}}\\ P_{\mathrm{sn}},& G_{\mathrm{bt}}\&GreaterEqual;G_{\mathrm{stf}}\end{array}\right.;---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit is kW; P _snfor the rated power of photovoltaic, be illustrated in the power that under standard test condition, unit light intensity can produce; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor a certain particular light intensity, under this intensity of illumination, photovoltaic is exerted oneself and is started from the non-linear linearity that becomes with the relation of intensity of illumination; G _btbe the real-time lighting intensity of t hour, unit is kW/m ²;

G _btreal-time lighting intensity by the sampling of the probability distribution statistical of historical light intensity is obtained;

S1-3 sets up battery model

$B_{t+1}=\left\{\begin{array}{cc}{B}_{\min },& B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\min }\\ B_t+{\mathrm{\&Delta;W}}_t& B_{\min }<B_t+{\mathrm{\&Delta;W}}_t\&le;B_{\max }\\ B_{\max },& B_{\max }<B_t+{\mathrm{\&Delta;W}}_t\end{array}\right.;---\left(5\right)$

In formula, Δ W _tfor t outside charge/discharge electricity amount of accumulator in the period, it equals to discharge and recharge the product of power and period t; B _tfor discharging and recharging the residual capacity of front accumulator, B _t=B _norm* Soc (t), wherein B _normfor battery rating, Soc (t) is the state-of-charge before discharging and recharging, B _t+1for discharging and recharging the residual capacity that finishes rear accumulator; B _min, B _maxbe respectively maximum, the minimum capacity of accumulator;

S1-4 sets up load model

L _t＝L _p×P _w×P _d×P _h(t)；

In formula, L _pfor year load peak, P _wfor the value in year-all load curves corresponding with t hour, P _dfor the value in the week-daily load curve corresponding with t hour, P _h(t) be the value in the day-hour load curve corresponding with t hour; Load value when Lt is t.

2. the simulation evaluation method of microgrid electric power system reliability according to claim 1, is characterized in that: described step S2 comprises following sub-step:

S2-1 asks the front average operating time of component failure and mean repair time

In micro-grid system, most elements are repairable elements, and the cyclic process of " operation-stop transport-operation " that its state variation situation can be by stable state is simulated, and the parameter of element meets following relational expression:

$\left\{\begin{array}{c}\mathrm{MTTR}=1/\mathrm{\&lambda;}\\ \mathrm{MTTF}=1/\mathrm{\&mu;}\end{array}\right.;$

In formula, MTTF is average operating time before component failure, the mean repair time that MTTR is element, is respectively the average of TTF and TTR; λ is crash rate, i.e. in Failure count/year, μ is repair rate, repairs number of times/year, and TTF is working time before losing efficacy, i.e. time between failures, and TTR is repair time, i.e. the out-of-service time;

S2-2 asks the time of front working time of the inefficacy of element and reparation

When micro-grid system is carried out to fail-safe analysis, it has been generally acknowledged that the next fault of element and fault are last time irrelevant, element fault has without memory; Therefore, think the equal obeys index distribution of time of working time and reparation before the inefficacy of element, its probability density function is:

$\left\{\begin{array}{c}f\left(t\right)={\mathrm{\&lambda;e}}^{-\mathrm{\&lambda;t}}\\ g\left(t\right)={\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, f (t) represents the probability that element breaks down constantly at t; G (t) represents the probability that element has been repaired constantly at t; F (t) with the probability distribution function of g (t) is:

$\left\{\begin{array}{c}F\left(t\right)={1-e}^{-\mathrm{\&lambda;t}}\\ G\left(t\right)={1-\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.;$

In formula, F (t) represents that element fault moment is less than the probability of t; G (t) represents that element reparation is less than the probability of t constantly;

Above formula is slightly done to change:

$\left\{\begin{array}{c}F\text{'}\left(t\right)={1-F=e}^{-\mathrm{\&lambda;t}}\\ G\text{'}\left(t\right)={1-G=\mathrm{\&mu;e}}^{-\mathrm{\&mu;t}}\end{array}\right.$

Wherein, F ' (t) represents the probability that the time between failures of element is t; G ' (t) represents the probability that repair time of element is t;

So, F ' (t) and G ' (t) be the number in interval [0,1]; Therefore, be positioned at the mode of the random number between [0,1] by generation, conversely the time between failures of sampling element and repair time, its sampling formula is:

$\left\{\begin{array}{c}\mathrm{TTF}=-\frac{1}{\mathrm{\&lambda;}}1n{R}_1\\ \mathrm{TTR}=-\frac{1}{\mathrm{\&mu;}}1n{R}_2\end{array}\right.;$

In formula, R ₁, R ₂be equally distributed random number between [0,1]; Adopt said method, to TTF and TTR difference alternate sampling.

3. the simulation evaluation method of microgrid electric power system reliability according to claim 2, is characterized in that: described step S3 comprises following sub-step:

S3-1 initialization simulated clock simulation clock is 0, produces at random m the random number between 0-1, according to the crash rate parameter lambda in each element state model try to achieve m non-failure operation time TTF, TTF _ithe TTF that represents i element;

S3-2 finds out minimum TTF _i;

S3-3 produces a random number to i element, and according to its repair rate parameter, μ tries to achieve fault correction time TTR _i;

S3-4 reads FMEA table, the load point affecting while searching element i fault, record these dead electricity load point frequency of power cut, power off time, lack the information such as delivery;

S3-5 produces a new random number, is translated into the running time T TF that element i is new _i';

S3-6 judges that whether simulated clock simulation clock is across year, not across the load point power failure information of record being added in year then in load point reliability index; If across year, adopt across year formula, calculate load point reliability index and the Reliability Index of this year;

Judging whether simulated clock simulation clock has been advanced to meets the required time span of Evaluation accuracy, if reach, performs step S3-7, does not reach and returns to step S3-2;

S3-7 simulation process finishes, and adds up the load point in each simulation year and the reliability index of system; The reliability index average of S3-8 and then calculating whole system.