CN108152673A - A kind of active distribution network failure Hierarchical Location method using multi-source data - Google Patents

Abstract

A kind of active distribution network failure Hierarchical Location method using multi-source data, is divided into prime area layer positioning and exact layer positions two levels.The information for comprehensively utilizing the breaker of the fault moment of upload, block switch and interconnection switch monitoring terminal FTU first carries out fault section area level positioning, then in the range of determining fault section, carry out line layer positioning, it is calculated using two end node electric quantity information of fault section, obtains accurate abort situation.Pass through verification, the results showed that the method for the present invention can not be influenced by fault type and position, quickly realize fault location；Used fault location principle is simple and reliable, has relatively strong anti-transition resistance ability；Sample frequency is of less demanding, while can improve fault location precision, therefore with higher actual application value by improving sample frequency.

Description

A Hierarchical Fault Location Method for Active Distribution Network Using Multi-source Data

technical field

The invention belongs to the technical field of electric power, relates to electric power system relay protection, and is a layered positioning method for active distribution network faults using multi-source data.

Background technique

With the development of modern power system, the network of distribution network is becoming more and more complex, and the capacity and voltage level are constantly increasing. The safe and normal operation of distribution network has become a key link in serving grid users. In the operation monitoring of distribution network, it is the key to improve the reliability of power supply of distribution network to quickly realize accurate fault location and distance measurement after line fault.

When a phase-to-phase short circuit fault occurs on the line, it will be accompanied by an overcurrent phenomenon, which is easy to realize fault identification. The single-phase-to-ground fault is the most common type of fault in the distribution network, and the fault location of the distribution network is mainly aimed at the accurate location of the single-phase-to-ground fault ^[1] . At present, the common methods of fault location in distribution network are: 1) Impedance-based fault location method, which measures the fault distance at the end of the line ^[2]-[3] . Impedance method is widely used in practical applications, and its advantage is that it is relatively simple and reliable , the disadvantage is that the ranging accuracy is not high. The traditional RL model algorithm is a single-ended fault location algorithm. In the distribution network application, the traditional RL model algorithm cannot eliminate the coupling information of the double-ended current in the case of transition resistance. , can not achieve accurate positioning; 2) the fault location method based on traveling waves, the fault distance is calculated by the traveling wave generated by the fault between the fault point and the measurement terminal round-trip time ^[4] ; 3) the fault location method based on signal injection, Fault point tracing is realized by injecting signals into the system ^[5] . Among them, the fault location method based on the impedance method is simple in principle and has been widely applied and researched. However, the fault location method based on the impedance method still has defects and needs to be further improved.

references

[1] Ji Tao, Sun Tongjing, Xue Yongduan, et al. Current status and prospect of distribution network fault location technology [J]. Power System Protection and Control, 2005,33(24):32-37.

[2] Huang Yanquan, Xiao Jian, Li Jin, et al. Preliminary study on the application of least squares method in distance protection [J]. Power System Protection and Control, 2004, 32(7): 17-20.

[3] Liu Jiajun, Li Chunju, Li Wenling. Research on Impedance Method Using Fault Component Current in Fault Location [J]. Electrified Railway, 2003(5):1-4.

[4] Dong Xinzhou, Ge Yaozhong, Xu Bingyin. Research on transmission line fault location using transient current traveling waves [J]. Chinese Journal of Electrical Engineering, 1999 (4): 76-80.

[5] Wang Xinchao, Sang Zaizhong. A New Method of Fault Location Based on "S Injection Method" [J]. Power System Protection and Control, 2001, 29(7): 9-12.

Contents of the invention

The problem to be solved by the present invention is that the impedance-based fault location method is widely used in distribution network fault location, but its location accuracy cannot meet the demand and needs to be improved. The invention proposes a new layered positioning method for active distribution network faults using multi-source data.

The technical solution of the present invention is: a hierarchical positioning method for active distribution network faults using multi-source data, including the following steps:

Step 1: After a fault occurs on the line, first use multi-source data to construct a fault description matrix to initially locate the fault location, first mark the circuit breaker, section switch and tie switch as switch nodes, and construct an N×N order switch network description matrix D. When the line fails, update the switch network description matrix D to obtain the fault description matrix F, and determine the faulty section;

Step 2: According to the information of the fault section obtained in step 1, in the initially determined fault section, the voltage and current information of the nodes m and n at both ends of the fault section are listed, and the equivalent resistance of the fault point at the m terminal is calculated and equivalent inductance:

Assuming that a short-circuit fault of transition resistance R _g occurs on the line between node m and node n, the establishment formula is:

In the formula, R _m and L _m are the equivalent resistance and inductance of the line at the distance m terminal where the fault occurs, respectively; R _n and L _n are the equivalent resistance and inductance of the line at the distance n terminal where the fault occurs; R _l and L _l line The effective total resistance and inductance value; u _m and u _n are the ground voltage at both ends of the line; i _m and i _n are the current at both ends of the line; and is the differential value of the current at both ends of the line; if _is the current at the fault point;

Arranging formula (1), we get:

Namely: Δu _mn ＝R _m i _f +L _m D _mn -U _n

where Δu _mn =u _m -u _n ; if _f = _im +i _n ;

At two moments _t1 and _t2 , respectively measure the voltage u, current i and current differential value at both ends of the ground resulting in two independent equations:

Time t ₁ :

Time t ₂ :

Simultaneous formulas (3) and (4) get:

In the formula: use the difference to approximate the differential calculation, and select the intermediate value of two adjacent sampling moments as the measured value at this moment, then:

Among them, T _s is the sampling interval time; i _m (h), i _m (h+1), i _m (h+2) are the hth, h+1 and h+2 sampling points of node m current respectively Values, i _n (h), i _n (h+1), i _n (h+2) are the values of the hth, h+1 and h+2 sampling points of the node n current respectively;

If there is no fault on the line, the simplification of formulas (5) and (6) will make the numerator and denominator both zero, and the fault distance calculation should not be performed. Therefore, the fault distance calculation criterion is set as follows:

|i _f1 D _mn2 -i _f2 D _mn1 |＞K _set (7)

Take K _set as 0.1, if there is no fault on the line, that is, the calculation criterion of dynamic fault distance is not satisfied, then the fault distance calculation will not be performed;

Step 3: Calculate the distance between the fault point and node m by using the equivalent resistance R _m and inductance L _m of the node m calculated in step 2 to obtain the distance between the fault point and node m, and use the average value of the fault distances calculated separately as the final fault distance calculation value.

Further, in step 3, the fault distances are obtained by R _m and L _m respectively, and the average value of the fault distances calculated separately is used as the final fault distance calculation value:

In the formula, R ₁ and L ₁ are the inductance value and resistance value of the unit length of the line respectively. The 5ms data window is selected to process the data from the fault section information. The data window slides with the sampling point. In a data window, the continuously obtained If the calculated value s of the fault location satisfies that the relative error is continuously less than 1%, it is judged that the calculated value is convergent, and it is taken as the final fault distance.

In order to solve the fault location problem of active distribution network lines, the present invention proposes a layered location method for active distribution network faults using multi-source data. On the basis of the traditional fault location method, the present invention comprehensively uses multi-dimensional data such as the information of the circuit breaker, section switch and tie switch monitoring terminal of the power grid, and at the same time improves the R-L model algorithm in the traditional single-ended fault location , so that it can not only be used for double-ended fault location, but also accurate. The method of the present invention is divided into two levels of initial area layer positioning and precise line layer positioning. First, the fault description matrix is constructed by using the uploaded information of the circuit breaker, section switch and tie switch monitoring terminal at the time of the fault to locate the area layer of the fault section. Then, within the range of the determined fault section, the line layer location is carried out, and the electrical quantity information of the nodes at both ends of the fault section is used for calculation to obtain an accurate fault location. The method of the invention can quickly realize fault location without being affected by the type and location of the fault; the adopted fault location principle is simple and reliable, and has strong resistance to transition resistance. The method of the invention has low requirements on the sampling frequency, and can improve fault location accuracy by increasing the sampling frequency, so it has good practical value.

Description of drawings

Fig. 1 is a 9-node active distribution network system according to an embodiment of the present invention.

Fig. 2 is a switch network diagram of a 9-node active and active distribution network system according to an embodiment of the present invention.

Fig. 3 is a single-phase-to-ground short-circuit system model of an embodiment of the present invention.

Detailed ways

On the basis of the traditional fault location method, the present invention comprehensively uses multi-source data such as circuit breaker, section switch and tie switch monitoring terminal FTU (Feeder terminal unit, feeder terminal equipment) information in the power grid, and simultaneously analyzes the traditional R-L model algorithm Improvements are made, and a hierarchical location method for active distribution network faults using multi-source data is proposed. The method is divided into two levels: initial area-level positioning and precise line-level positioning. Firstly, the fault description matrix is constructed by using the uploaded monitoring information of the circuit breaker, section switch and tie switch at the time of the fault to determine the scope of the fault section, and then within the scope of the determined fault section, the improved The R-L model algorithm calculates the exact location of the faulty line. Through verification, the method is not affected by the type and location of the fault, has a strong ability to resist transition resistance, and can quickly and accurately locate the fault.

The implementation method of the present invention is described below.

Step 1: After a fault occurs on the line, first use multi-source data such as circuit breaker, section switch and tie switch monitoring terminal FTU information to construct a fault description matrix for initial location of the fault location. Based on the topological structure of the active distribution network, circuit breakers, section switches and tie switches are marked as switch nodes, and a switch network description matrix D is formed. First, set the positive direction of the network, and stipulate that: the single power supply network is the direction of line power flow; the multi-power supply network is the direction of line power flow under the assumption that a single power supply is supplied. In a network of N nodes, construct an N×N-order switch description matrix D. For switch node i, if switch node j is connected to it, and the forward direction of the network is that switch node i points to switch node j, d _ij = 1; otherwise d _ij ＝0. That is to say:

When a line fault occurs, if the direction of the fault current flowing through the switch node i is consistent with the set positive direction of the network, the node FTU will upload a signal "1" to the control center, and at the same time modify the switch network description matrix D, and correct the value of d _ii is 1; otherwise, if the direction of the fault current flowing through the switch node i is opposite to the positive direction of the network, or no fault current flows, then d _ii maintains the original value. After the switch network description matrix D is corrected, the fault description matrix F is obtained.

According to the fault description matrix F, the regional layer fault location is carried out. If the switch node i satisfies f _ii =1, the fault region judgment principle is as follows:

(1) For all switching nodes j (j=1,2,3,...N, j≠i) satisfying f _ij =1, and satisfying f _jj =0 at the same time, the fault occurs in switching node i and switching node j paragraph.

(2) If all switching nodes j (j=1, 2, 3, ... N, j≠i) satisfy f _ij =0, then the fault occurs at the end of switching node i.

Step 2: Use multi-source data such as circuit breaker, section switch and tie switch to monitor terminal FTU information. After the initial location of the fault location is completed, the information of the fault section can be obtained. In the initially determined fault section, according to the voltage and current information of the nodes at both ends of the fault section, the column formula can be carried out. For example, when a short-circuit fault occurs on the line between node m and node n through the transition resistance _Rg , the formula is established as:

In the formula, R _m and L _m are the equivalent resistance and inductance of the line at the distance m terminal where the fault occurs, respectively; R _n and L _n are the equivalent resistance and inductance of the line at the distance n terminal where the fault occurs; R _l and L _l line The effective total resistance and inductance value; u _m and u _n are the ground voltage at both ends of the line; i _m and i _n are the current at both ends of the line; and is the differential value of the current at both ends of the line; if _is the current at the fault point.

Arranging the formula, we can get:

That is: Δu _mn ＝R _m i _f +Z _m D _mn -U _n

where Δu _mn =u _m -u _n ; if _f = _im +i _n ;

Measure u _, i _and Two independent equations can be obtained:

Then it can be calculated by simultaneous formulas (4) and (5). get:

In the formula: use the difference to approximate the differential calculation, and select the intermediate value of two adjacent sampling moments as the measured value at this moment, then:

Among them, T _s is the sampling interval time; i _m (h), i _m (h+1), i _m (h+2) are the hth, h+1 and h+2 sampling points of node m current respectively The values of i _n (h), i _n (h+1), i _n (h+2) are the values of the hth, h+1 and h+2 sampling points of the node n current respectively.

If there is no fault on the line, the simplification of formulas (6) and (7) will make the numerator and denominator both zero, and the fault distance calculation should not be performed. Therefore, the fault distance calculation criterion is set as follows:

|i _f1 D _mn2 -i _f2 D _mn1 |＞K _srt (8)

Take K _set as 0.1, if there is no fault on the line and the criterion for calculating the dynamic fault distance is not met, then the fault distance calculation will not be performed.

Step 3: Use step 2 to calculate the equivalent resistance and reactance value of the node m at a distance from the fault location. Therefore, the fault distance is obtained by R _m and L _m respectively, and the average value of the fault distances calculated separately is used as the final fault distance calculation value:

In the formula, R ₁ and L ₁ are the inductance value and resistance value per unit length of the line, respectively. A 5ms data window is selected for data processing, and the data window slides along with the sampling points. In a data window, if the continuously obtained calculated value s of the fault location satisfies that the relative error is continuously less than 1%, then it is judged that the calculated value converges, and this is taken as the final fault distance.

The content of the present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.

Figure 1 shows the structure of the 9-node active and active distribution network system. The system has two distributed generation DGs and is connected to the main grid UG (Utility Grid) at the same time. Among them, CB1-CB7 represent circuit breakers, S1-S3 represent section switches, and FTUs are installed on both circuit breakers and section switches, T1-T3 represent main transformers, and T4-T6 represent distribution transformers. The circuit breaker and section switch are marked and numbered as switch nodes in turn, and the switch network diagram can be obtained as shown in Figure 2. Assume that the positive direction of the network is the power flow direction when DG1 supplies power alone, which is the direction of the solid arrow. From this, the switch network description matrix D can be obtained as:

Set A-phase line grounding fault F1 on line line4 10km away from node 7, the transition resistance is 1Ω, and the fault occurs at t=0.5s. As shown in Figure 2, according to the principle that the direction of the fault current is consistent with the positive direction of the network, it can be obtained that d ₂₂ =d ₆₆ =1, and then the matrix D is corrected to obtain the fault description matrix F as:

According to the fault area judgment principle (1):

f _6,6 =1, f _6,10 =1, f _10,10 =0, then the fault area is between switch node 4 and switch node 10, and corresponding to the system structure diagram is between node 7 and node 8.

When a fault occurs, the power information of each node is uploaded to the central computing platform, and the initial location of the fault has determined that the fault interval is between node 7 and node 8, so that the entire system can be equivalent to a double-terminal power supply model as shown in Figure 3 . Using the current and voltage information of nodes 7 and 8, the fault location can be calculated.

According to the system model, the time-domain differential equation of phase A voltage at terminal m is listed as:

Among them, K _R ＝(R ₀ -R ₁ )/3R ₁ , K _L ＝(L ₀ -L ₁ )/3L ₁ are the zero-sequence compensation coefficients of line resistance and inductance respectively; R ₀ and R ₁ are line units Positive-sequence and zero-sequence resistance, L ₀ and L ₁ are positive-sequence and zero-sequence inductance of line units respectively; u _ma , i _ma , i _m0 are bus measurement terminal voltage, current and zero-sequence current.

In the same way, the time-domain differential equation of the n-terminal A-phase voltage can be obtained as:

Among them, R _m and L _m are the equivalent resistance and inductance of the line at the distance m terminal where the fault occurs, respectively; R _n and L _n are the equivalent resistance and inductance of the line at the distance n terminal where the fault occurs; R _l and L _l line equivalent The total resistance and inductance; u _na , _ina , in _n0 are the measured terminal voltage, current and zero-sequence current of the A-phase bus in the n-terminal fault state; _if is the fault point current.

Measure _u , i _and Two independent equations can be obtained:

Then R _m and L _m can be obtained through simultaneous calculation, and the average value of the fault distance obtained by R _m and L _m respectively is used as the final fault distance calculation value:

In the formula, R ₁ and L ₁ are the positive sequence resistance value and inductance value per unit length of the line, respectively. The present invention selects a 5ms data window for data processing, and the data window slides along with the sampling points. In a data window, if the continuously obtained calculated value s of the fault location satisfies that the relative error is continuously less than 1%, then it is judged that the calculated value is converged, and this is taken as the final fault distance.

With the development of modern power systems, the safe and normal operation of distribution networks has become a key link in serving grid users. After a line fault occurs in the distribution network, fast and accurate fault location and distance measurement is the key to improving the reliability of power supply in the distribution network. On the basis of the traditional fault location method, the present invention comprehensively uses multi-dimensional data such as the information of the circuit breaker, section switch and tie switch monitoring terminal of the power grid, and improves the traditional R-L model algorithm at the same time, and proposes a method using multiple Active distribution network fault hierarchical location method based on source data. The method is divided into two levels: initial regional layer positioning and precise line layer positioning. First, the fault description matrix is constructed by using the uploaded information of the circuit breaker, section switch and tie switch monitoring terminal at the time of the fault to locate the fault section regional layer, and then Within the scope of the determined fault section, the line layer location is carried out, and the electrical quantity information of the nodes at both ends of the fault section is used for calculation to obtain an accurate fault location. It has been verified that this method can quickly realize fault location without being affected by the type and location of the fault; the fault location principle adopted is simple and reliable, and has a strong ability to resist transition resistance; Positioning accuracy, so it has high practical application value.

Claims (3)

1. a kind of active distribution network failure Hierarchical Location method using multi-source data, it is characterized in that including the following steps：

Step 1：After line failure, failure-description matrix is constructed first with multi-source data, carries out the initial of abort situation Breaker, block switch and interconnection switch are marked as switching node by positioning first, build N × N rank switching network Description Matrixes D, works as line failure, and update switching network Description Matrix D obtains failure-description matrix F, judges fault section；

Step 2：The fault section information obtained according to step 1, in originally determined fault section, according to fault section both ends The voltage and current information of node m, n calculate the equivalent resistance and equivalent inductance at trouble point distance m ends into determinant：

It is located at circuit between node m and node n and transition resistance R occurs_gShort trouble, establishing formula is：

In formula, R_mAnd L_mThe respectively line equivalent resistance and inductance at failure point distance m ends, R_nAnd L_nRespectively failure occurs The line equivalent resistance and inductance at place distance n ends；R_lAnd L_lLine equivalent all-in resistance and inductance value；u_mAnd u_nFor circuit both ends pair Ground voltage；i_mAnd i_nFor circuit both ends electric current；WithFor circuit both ends current differential value；i_fFor current in the fault point；

Formula (1) is arranged, is obtained：

I.e.：△u_mn=R_mi_f+L_mD_mn-U_n

Wherein △ u_mn=u_m-u_n；i_f=i_m+i_n；

In two moment t₁And t₂Voltage-to-ground u, electric current i and the current differential value at both ends are measured respectivelyObtain two it is independent Equation：

t₁Moment：

t₂Moment：

Simultaneous formula (3) and (4) obtain：

In formula：It is calculated using difference come approximate differential, while chooses the median of two adjacent sampling instances as the moment Measured value, then have：

Wherein, T_sFor sampling interval duration；i_m(h), i_m(h+1), i_m(h+2) be respectively node m electric currents h, h+1 and h+2 The value of sampled point, i_n(h), i_n(h+1), i_n(h+2) be respectively node n electric currents h, h+1 and h+2 sampled points value；

If circuit does not break down, formula (5) (6), which carries out abbreviation, can so that molecule and denominator are zero simultaneously, should not carry out failure Distance calculates, therefore fault distance is set to calculate criterion：

Take K_setIt is 0.1, if circuit does not break down, that is, is unsatisfactory for dynamic fault distance and calculates criterion, then without fault distance meter It calculates；

Step 3：The equivalent resistance R of abort situation distance m nodes is calculated using step 2_mWith inductance L_mTrouble point is acquired respectively With the distance of node m, and the average value of the fault distance to be calculated respectively is as final fault distance calculated value.

2. a kind of active distribution network failure Hierarchical Location method using multi-source data according to claim 1, feature It is in step 3, with R_mAnd L_mAcquire fault distance respectively, using the average value of fault distance that is calculated respectively as final therefore Barrier is apart from calculated value：

In formula, R₁And L₁Respectively circuit unit length inductance value and resistance value choose 5ms data windows to by fault section information Data processing is carried out, data window is slided with sampled point, and in a data window, the abort situation calculated value s being continuously available expires Sufficient relative error is continuously less than 1%, then judges that calculated value is restrained, as final therefore distance.

3. a kind of active distribution network failure Hierarchical Location method using multi-source data according to claim 1, feature It is that step 1 is specially：

Based on topological structure in active distribution network, breaker, block switch and interconnection switch are marked as switching node, set net Network positive direction in the network of N number of node, builds N × N rank switching network Description Matrix D, for switching node i, if joint Point j is attached thereto, and network positive direction is switching node i directing switch node j, then d_ij=1；Otherwise d_ij=0, that is, have：

Work as line failure, if the direction that switching node i flows through fault current is consistent with the network positive direction set, this section The FTU of point corrects switching network Description Matrix D to control centre's up-delivering signal " 1 ", by d_iiValue is modified to 1；If it otherwise opens Artis i flow through fault current direction with network positive direction on the contrary, no fault current flow through, then d_iiInitial value is kept, is opened It closes after the completion of network Description Matrix D is corrected and obtains failure-description matrix F, area level failure is carried out according to failure-description matrix F Positioning, if switching node i meets f_ii=1, fault section detection principle is：

1) meet f for all_ij=1 switching node j, j=1,2,3 ... N, j ≠ i, if meeting f simultaneously_jj=0, then failure send out Life is on switching node i and switching node j sections；

2) if all switching node j are satisfied by f_ij=0, then failure be happened at the end of switching node i.