CN106124933A - A kind of power system failure diagnostic method based on input nonlinearities method of equal value - Google Patents

Abstract

The invention discloses a kind of electric power system fault determination methods based on input nonlinearities method of equal value, comprise the steps: S1, be rationally layered whole power system, and divide group on the basis of every layer, group divides again subgroup, calculates the power of equal value of each group；S2, build the input nonlinearities observer of equal value of each group；S3, obtain the input nonlinearities threshold value of equal value of group；S4, from the beginning of ground floor, the most down observe the situation of change of the input nonlinearities observer of equal value of group, if occurring, its value exceeds the situation of input nonlinearities threshold value of equal value in certain time period, can determine whether to break down, bottom group down detected, i.e. determine the concrete node of fault, realize location of fault is judged, if the situation beyond threshold value does not occurs, then continue monitoring, to there is not the group beyond threshold value, its subgroup need not be detected again.Present invention, avoiding the defect that tradition corresponding failure diagnostic method is caused based on protection device and chopper information, failure diagnosis time is short and precision is high.

Description

A kind of power system failure diagnostic method based on input nonlinearities method of equal value

Technical field

The present invention relates to a kind of electric power system fault determination methods based on input nonlinearities method of equal value, more precisely, relate to And a kind of directly utilize the method that in power system, basic electric parameters carries out breakdown judge.

Background technology

Power system failure diagnostic is the important technology that power system maintains stability, judges system fast and accurately Present in fault be after a step fault is made the basis of respective handling, namely the requirement of fault diagnosis.

Power system failure diagnostic mainly has expert knowledge system, Petri network, genetic algorithm, multi-agent technology etc. at present Deng some correlation methods, they depend on protection device and the information of chopper in system, have two major defects:

1), in the case of sensor fault, owing to can not get complete information, the accurate of fault diagnosis can be had a strong impact on Property；

2), in above method research, need system whole-network is modeled, and when network size is bigger, system mould Shape parameter will become huge, and Diagnostic Time is elongated and precise decreasing.

As can be seen here, prior art there is also certain deficiency.

Summary of the invention

In view of this, in order to solve power system failure diagnostic length in prior art and the low technical problem of precision, this A kind of electric power system fault determination methods based on input nonlinearities method of equal value of bright proposition, relies only on basic electric parameters in power system Just can carry out the input nonlinearities hierarchical fault diagnosis method of equal value of fault diagnosis, the raising to power system failure diagnostic has important Meaning.

The present invention solves the problems referred to above by techniques below means:

A kind of electric power system fault determination methods based on input nonlinearities method of equal value, comprises the steps:

S1, being rationally layered whole power system, and divide group on the basis of every layer, group divides again subgroup, meter Calculate the power of equal value of each group；

S2, build the input nonlinearities observer of equal value of each group；

S3, obtain the input nonlinearities threshold value of equal value of group；

S4, from the beginning of ground floor, the most down observe the situation of change of the input nonlinearities observer of equal value of group, if there is it It is worth in certain time period beyond the situation of input nonlinearities threshold value of equal value, can determine whether to break down, bottom group down detected, the most really Determine the concrete node of fault, it is achieved location of fault is judged, if the situation beyond threshold value does not occurs, then continue monitoring, right The group beyond threshold value do not occur, its subgroup need not be detected again.

Further, also include:

S5, to finally determining the minimum subgroup of fault, take its state variable and be observed, and it is carried out corresponding differential Algorithm process, determines whether current phase and amplitude break down, it is achieved judge the type of fault.

Further, in step S1, whole power system is layered by network topology structure and the logic of each node Relation determines, the power of equal value of group is realized by forward-backward sweep method Load flow calculation.

Further, in step S2, input nonlinearities observer parameter of equal value is calculated as follows:

$\begin{array}{cc}{A}_j^{\left(k\right)}=-\frac{r_1^{\left(k\right)}+R_1^{\left(k\right)}}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}& B_j^{\left(k\right)}=\frac{1}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}\end{array}$

$\begin{array}{cc}{C}_j^{\left(k\right)}=R_1^{\left(k\right)}-\frac{L_{h1}^{\left(k\right)}(r_1^{\left(k\right)}+R_1^{\left(k\right)})}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}& D_j^{\left(k\right)}=\frac{L_{h1}^{\left(k\right)}}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}\end{array}---\left(1\right)$

Wherein,WithIt is node represented by kth layer j group to the equivalent resistance of circuit between a node on it and electricity It is anti-,WithIt is this group of equivalent resistance and reactance；

Further, in step S2, input nonlinearities observer gain value of equal value is determined by quadratic performance optimization.

Further, in step S2, input nonlinearities observer gain value of equal value is determined by quadratic performance optimization, tool Body is as follows:

For determining input nonlinearities observer gain value K of equal value, performance index function uses Quadratic Function Optimization, namely uses error Integrated square

$J={\∫}_0^{\mathrm{\∞}}{e}^2\left(t\right)dt={\∫}_0^{\mathrm{\∞}}\[y_r\left(t\right)-y\left(t\right)\]dt---\left(2\right)$

Evaluate the quality of system Control platform, owing to controlled quentity controlled variable is constrained, in integrand consider increase by one with Controlling the penalty term that power is relevant, rewriteeing its performance index function is

$J={\∫}_0^{\mathrm{\∞}}\[k_1{e}^2\left(t\right)+k_2{e}^2\left(t\right)\]dt---\left(3\right)$

Infinite Time standing state actuator: set state equation and the quadratic performance index of Linear Time Invariant controlled system It is respectively

$\begin{array}{cc}\stackrel{\·}{x}\left(t\right)=Ax\left(t\right)+Bu\left(t\right)& x\left(0\right)=x_0\end{array}$

$J={\∫}_0^{\mathrm{\∞}}\[x^T\left(t\right)Qx\left(t\right)+u^T\left(t\right)Ru\left(t\right)\]dt---\left(4\right)$

$J={\∫}_0^{\mathrm{\∞}}\[x^T\left(t\right)Qx\left(t\right)+u^T\left(t\right)Ru\left(t\right)\]dt$

In formula: x (t) is the state variable of n dimension, u (t) is the dominant vector of p dimension, and A, B, Q and R are then to have suitable dimension Constant matrices；And weighted matrix R=R^T＞ 0, Q=Q^T＞ 0 or Q=Q^T>=0, and { the complete Observable of A, D}, D is among these Make DD^TThe Arbitrary Matrix that=Q sets up, determines an optimum control u from allowing control^*T (), so that performance indications minimize；

In Matlab, function lqr () is used to solve Infinite Time standing state regulator problem, i.e.

[K, P, E]=lqr (A, B, Q, R) (5)

Simply enter coefficient matrices A, B and the quadratic performance index of linear controlled system state equation weighting matrices Q, R, just can try to achieve the state feedback matrix K of optimum output adjusting system, input nonlinearities observer gain value K i.e. of equal value.

Further, in step S3, input nonlinearities threshold value of equal value utilizes the greatest measure of power system load curve to pass through The output valve of input nonlinearities observer of equal value determines.

Further, in step S5, differential algorithm includes that parity price input nonlinearities observer waveform peak is carried out at differential Reason, current amplitude is calculated by the maximin of load, and current phase is calculated by power factor.

Compared with prior art, present invention electric power system fault determination methods based on input nonlinearities method of equal value is big by one Power system network, be divided into multistage multilayer, by the way of successively launching, to diagnosing malfunction and location, the present invention's Fault diagnosis technology directly uses the direct electric parameters in power system to carry out model construction, it is to avoid the diagnosis of traditional corresponding failure The defect that method is caused based on protection device and chopper information, fault diagnosis is accurate, and Diagnostic Time is short and precision is high.

Accompanying drawing explanation

For the technical scheme being illustrated more clearly that in the embodiment of the present invention, in embodiment being described below required for make Accompanying drawing be briefly described, it should be apparent that, below describe in accompanying drawing be only some embodiments of the present invention, for From the point of view of those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to obtain other according to these accompanying drawings Accompanying drawing.

Fig. 1 is for the structure chart of IEEE33 node power distribution network used by detailed description of the invention is described；

Fig. 2 is the flow chart of present invention electric power system fault determination methods based on input nonlinearities method of equal value；

Fig. 3 is IEEE33 network hierarchy illustraton of model；

Fig. 4 is input nonlinearities Observer Structure figure of equal value.

Detailed description of the invention

Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, below in conjunction with accompanying drawing with concrete Embodiment technical scheme is described in detail.It is pointed out that described embodiment is only this Bright a part of embodiment rather than whole embodiments, based on the embodiment in the present invention, those of ordinary skill in the art are not having Have and make the every other embodiment obtained under creative work premise, broadly fall into the scope of protection of the invention.

Embodiment

For illustrating that the present invention's is embodied as details, as a example by IEEE33 node power distribution network (as shown in Figure 1), illustrate The present invention specifically carries out the process of fault diagnosis.

As in figure 2 it is shown, a kind of electric power system fault determination methods based on input nonlinearities method of equal value, comprise the steps:

S1, being rationally layered whole power system, and divide group on the basis of every layer, group divides again subgroup, as Shown in Fig. 3, calculate the power of equal value of each group；

Whole power system is layered and is determined by the logical relation of network topology structure and each node, the merit of equal value of group Rate is realized by forward-backward sweep method Load flow calculation；

S2, build the input nonlinearities observer of equal value of each group as shown in Figure 4；

Input nonlinearities observer parameter of equal value is calculated as follows:

$\begin{array}{cc}{A}_j^{\left(k\right)}=-\frac{r_1^{\left(k\right)}+R_1^{\left(k\right)}}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}& B_j^{\left(k\right)}=\frac{1}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}\end{array}$

$\begin{array}{cc}{C}_j^{\left(k\right)}=R_1^{\left(k\right)}-\frac{L_{h1}^{\left(k\right)}(r_1^{\left(k\right)}+R_1^{\left(k\right)})}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}& D_j^{\left(k\right)}=\frac{L_{h1}^{\left(k\right)}}{L_{f1}^{\left(k\right)}+L_{h1}^{\left(k\right)}}\end{array}---\left(1\right)$

Wherein,WithIt is node represented by kth layer j group to the equivalent resistance of circuit between a node on it and electricity It is anti-,WithIt is this group of equivalent resistance and reactance；

Input nonlinearities observer gain value of equal value is determined by quadratic performance optimization, specific as follows:

For determining input nonlinearities observer gain value K of equal value, within the system, performance index function uses quadratic form letter Number, namely with the integration of square-error

$J={\∫}_0^{\mathrm{\∞}}{e}^2\left(t\right)dt={\∫}_0^{\mathrm{\∞}}\[y_r\left(t\right)-y\left(t\right)\]dt---\left(2\right)$

Evaluate the quality of system Control platform, owing to controlled quentity controlled variable is constrained, in integrand consider increase by one with Controlling the penalty term that power is relevant, rewriteeing its performance index function is

$J={\∫}_0^{\mathrm{\∞}}\[k_1{e}^2\left(t\right)+k_2{e}^2\left(t\right)\]dt---\left(3\right)$

In Practical Project, optimal adjustment system should asymptotically stability, and its transient state quality should be excellent, Feedback gain matrix is also permanent, thus could constitute a permanent feedback system, and Infinite Time standing state regulates What device problem was provided most has control system to be proved to is exactly so one system；

Infinite Time standing state actuator: set state equation and the quadratic performance index of Linear Time Invariant controlled system It is respectively

$\begin{array}{cc}\stackrel{\·}{x}\left(t\right)=Ax\left(t\right)+Bu\left(t\right)& x\left(0\right)=x_0\end{array}$

$J={\∫}_0^{\mathrm{\∞}}\[x^T\left(t\right)Qx\left(t\right)+u^T\left(t\right)Ru\left(t\right)\]dt---\left(4\right)$

$J={\∫}_0^{\mathrm{\∞}}\[x^T\left(t\right)Qx\left(t\right)+u^T\left(t\right)Ru\left(t\right)\]dt$

In formula: x (t) is the state variable of n dimension, u (t) is the dominant vector of p dimension, and A, B, Q and R are then to have suitable dimension Constant matrices；And weighted matrix R=R^T＞ 0, Q=Q^T＞ 0 or Q=Q^T>=0, and { the complete Observable of A, D}, D is among these Make DD^TThe Arbitrary Matrix that=Q sets up, our target is to determine an optimum control u control from allowing^*(t), so that performance Index minimizes；

In Matlab, function lqr () is used to solve Infinite Time standing state regulator problem, i.e.

[K, P, E]=lqr (A, B, Q, R) (5)

Simply enter coefficient matrices A, B and the quadratic performance index of linear controlled system state equation weighting matrices Q, R, just can try to achieve the state feedback matrix K of optimum output adjusting system, input nonlinearities observer gain value K i.e. of equal value；

S3, obtain the input nonlinearities threshold value of equal value of group；

Input nonlinearities threshold value of equal value utilizes the greatest measure of power system load curve by input nonlinearities observer of equal value Output valve determine；

S4, from the beginning of ground floor, the most down observe the situation of change of the input nonlinearities observer of equal value of group, if there is it It is worth in certain time period beyond the situation of input nonlinearities threshold value of equal value, can determine whether to break down, bottom group down detected, the most really Determine the concrete node of fault, it is achieved location of fault is judged, if the situation beyond threshold value does not occurs, then continue monitoring, right The group beyond threshold value do not occur, its subgroup need not be detected again；

S5, to finally determining the minimum subgroup of fault, take its state variable and be observed, and it is carried out corresponding differential Algorithm process, determines whether current phase and amplitude break down, it is achieved judge the type of fault；

Differential algorithm includes state observer waveform peak is carried out differential process, and current amplitude is by the maximum of load Little value calculates, and current phase is calculated by power factor.

The present invention provides a kind of and does not relies on protection device and the method for diagnosing faults of chopper information in power system, master To be carried structure observer by input nonlinearities method of equal value and to carry out breakdown judge.Input nonlinearities method of equal value is the one in modern scientist field New hot technology, uses extensively in the controls.Input nonlinearities method of equal value is intended to illustrate to exist in classical feedback system One input nonlinearities of equal value, can have consistent effect with former interference to the output that system produces, and by the input of this equivalence is done The suppression disturbed is to make system more stable.Power system is considered as the control system of classics by the present invention, and by power system Fault is equivalent to the interference in control system, thus is estimated interference by the input nonlinearities of equal value in input nonlinearities method of equal value, Namely failure judgement.Due to the complexity of power system network, whole power system network is carried out complete modeling and exists tired Difficulty, the present invention proposes a kind of power system failure diagnostic model based on layered distribution type thought, and breakdown judge base just In this model.

According to the present invention, by a big power system network, it is divided into multistage multilayer, by the way of successively launching, right Diagnosing malfunction and location.Include that the calculating of system interior joint parameter, the determination of observer gain, input of equal value are done among these The a series of program of determination disturbing threshold value etc., the fault diagnosis technology of the present invention directly use in power system directly electrically Amount carries out model construction, it is to avoid traditional corresponding failure diagnostic method is based on lacking that protection device and chopper information are caused Falling into, fault diagnosis is accurate, and Diagnostic Time is short and precision is high.

Embodiment described above only have expressed the several embodiments of the present invention, and it describes more concrete and detailed, but also Therefore the restriction to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that, for those of ordinary skill in the art For, without departing from the inventive concept of the premise, it is also possible to make some deformation and improvement, these broadly fall into the guarantor of the present invention Protect scope.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (8)

1. an electric power system fault determination methods based on input nonlinearities method of equal value, it is characterised in that comprise the steps:

S1, being rationally layered whole power system, and divide group on the basis of every layer, group divides again subgroup, calculates every The power of equal value of individual group；

S2, build the input nonlinearities observer of equal value of each group；

S3, obtain the input nonlinearities threshold value of equal value of group；

S4, from the beginning of ground floor, the most down observe the situation of change of the input nonlinearities observer of equal value of group, if occurring, its value exists Certain time period, beyond the situation of input nonlinearities threshold value of equal value, can determine whether to break down, bottom group down detected, i.e. determine The concrete node of fault, it is achieved judge location of fault, if there is not the situation beyond threshold value, then continues monitoring, to not going out Now beyond the group of threshold value, its subgroup need not be detected again.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that Also include:

S5, to finally determining the minimum subgroup of fault, take its state variable and be observed, and it is carried out corresponding differential algorithm Process, determine whether current phase and amplitude break down, it is achieved the type of fault is judged.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that In step S1, whole power system is layered and is determined by the logical relation of network topology structure and each node, the equivalence of group Power is realized by forward-backward sweep method Load flow calculation.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that In step S2, input nonlinearities observer parameter of equal value is calculated as follows:

Wherein,WithIt is node represented by kth layer j group to the equivalent resistance of circuit between a node on it and reactance,WithIt is this group of equivalent resistance and reactance.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that In step S2, input nonlinearities observer gain value of equal value is determined by quadratic performance optimization.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that In step S2, input nonlinearities observer gain value of equal value is determined by quadratic performance optimization, specific as follows:

For determining input nonlinearities observer gain value K of equal value, performance index function uses Quadratic Function Optimization, namely uses square-error Integration

Evaluate the quality of system Control platform, owing to controlled quentity controlled variable is constrained, consider increase by one and control in integrand The penalty term that power is relevant, rewriteeing its performance index function is

Infinite Time standing state actuator: set state equation and the quadratic performance index difference of Linear Time Invariant controlled system For

X (0)=x₀

In formula: x (t) is the state variable of n dimension, u (t) is the dominant vector of p dimension, and A, B, Q and R are then have suitable dimension normal Matrix number；And weighted matrix R=R^T＞ 0, Q=Q^T＞ 0 or Q=Q^T>=0, and { the complete Observable of A, D}, D is to make among these DD^TThe Arbitrary Matrix that=Q sets up, determines an optimum control u from allowing control^*T (), so that performance indications minimize；

In Matlab, function lqr () is used to solve Infinite Time standing state regulator problem, i.e.

[K, P, E]=lqr (A, B, Q, R) (5)

Simply enter the weighting matrices Q of coefficient matrices A, B and quadratic performance index of linear controlled system state equation, R, just The state feedback matrix K of optimum output adjusting system, input nonlinearities observer gain value K i.e. of equal value can be tried to achieve.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 1, it is characterised in that In step S3, input nonlinearities threshold value of equal value utilizes the greatest measure of power system load curve by input nonlinearities observer of equal value Output valve determine.

Electric power system fault determination methods based on input nonlinearities method of equal value the most according to claim 2, it is characterised in that In step S5, differential algorithm includes that parity price input nonlinearities observer waveform peak carries out differential process, and current amplitude is by negative The maximin carried calculates, and current phase is calculated by power factor.