CN103995162A - Power distribution network large user real-time electricity larceny prevention method based on advanced measuring system - Google Patents

Abstract

The invention belongs to the technical field of electricity larceny prevention of electric system distribution lines, and particularly relates to a power distribution network large user real-time electricity larceny prevention method based on an advanced measuring system. The method includes the steps that firstly, measured data of all intelligent electric meters are read, then, load flow calculation is conducted on user voltages and power at the current moment, the theoretical line loss active power and the theoretical line loss reactive power of a feeder line and the calculation voltages of all nodes are acquired, the electricity larceny power of a user is calculated, the electricity larceny behavior of the user is judged, if electricity larceny exists, monitoring points of voltage deviation are ranked from large to small, electricity larceny point positioning and electricity larceny amount calculation are conducted through a point-by-point electricity larceny power iterative optimization calculation method, and then electricity larceny alarming information is sent out. The method has the advantages of being high in data measurement and calculation accuracy, good in synchronism, high in speed and accurate in positioning, electricity larceny points can be found in real time and electricity larceny power can be calculated in real time, treatment is conducted immediately after the electricity larceny points are found, and losses of power supply enterprises are reduced to the maximum degree.

Description

The real-time electricity anti-theft method of power distribution network large user based on senior measurement system

Technical field

The invention belongs to the anti-theft electricity technology field of electric system distribution line, relate in particular to a kind of real-time electricity anti-theft method of power distribution network large user based on senior measurement system.

Background technology

For a long time, stealing electricity phenomenon is the problem that electric power enterprise is extremely paid close attention to always, and particularly large user's stealing is brought tremendous economic loss to country and electric power enterprise, but also brings a series of potential safety hazard and social concern.General in the electricity sales amount of various places power supply enterprise, 10kV dedicated transformer large user accounts for more than 70%.Therefore, do a good job of it 10kV dedicated transformer large user's anti-electricity-theft control, just can compared with great dynamics improve the economic benefit of power supply enterprise.Common stealing side is as follows: undercurrent stealing gimmick, under-voltage stealing gimmick, the poor stealing gimmick that expands, phase shift stealing gimmick, nothing table stealing gimmick, high-tech stealing gimmick etc.For above problem, power supply enterprise also actively develops the technical method of the aspect of opposing electricity-stealing, although obtained certain effect, these methods people can not control the generation of stealing electricity phenomenon completely, discovery user's that particularly can not be real-time electricity filching behavior.

The intelligent grid of the performances such as construction has flexibly, cleans, safety, economy, close friend is the developing direction of following electrical network.As most important technical support module in intelligent grid, senior measurement system (Advanced Metering Infrastructure, AMI) in intelligent grid, taking on very important role, in intelligent grid, a lot of intelligent functions are implemented by AMI and complete.It utilize intercommunication system and can recording user the intelligent electric meter of detailed information on load, can obtain user with time target at times or the multiple variable of (or quasi real time) in real time, as information such as power consumption, need for electricity, voltage, electric currents.Therefore, AMI is a basic functional module of intelligent grid, also referred to as intelligence, measures system.With traditional electric distribution network data collection and supervisory system (Supervisory Control And Data Acquisition, SCADA) compare, AMI metric data management system (Meter Data Management System, MDMS) can not only provide the metric data that data precision and sample frequency are higher, can also provide by load real-time measurement data, this is that distribution SCADA is not available.Senior measurement architectural schematic as shown in Figure 1.

The enforcement of AMI and application have improved the measurement redundance of distribution system greatly, for solid foundation has been established in the enforcement of the real-time electricity anti-theft method of power distribution network large user.AMI is the basic link of intelligent grid, and major function is the electric information (with markers) that gathers each monitoring point, and makes power consumer set up contact with load, supports operation of power networks.In the information gathering at AMI, continuous data occupies most important status.These continuous datas had both comprised active energy and the reactive energy that common electric energy meter can measure, and comprised again the electric network state data such as voltage, electric current, power factor, forward and reverse power.Continuous data be the prerequisite that AMI and even other functions of intelligent grid are achieved accurately and reliably.On the other hand, the continuous data that AMI gathers should not become the dead data that are kept in database, and should be used.By the abundant communication of AMI and calculating storage resources, data are carried out to deep processing and excavation, realize its value.The present invention is the real-time electricity anti-theft method based on AMI.

Summary of the invention

For current anti-theft electricity technology means, fall behind, and the measured data of AMI utilize the problem that degree is not high, propose a kind of real-time electricity anti-theft method of power distribution network large user based on senior measurement system, mainly comprise:

Step 1, read the metric data of each intelligent electric meter of MDMS current time, mainly comprise: the user node active-power P that each large user's distribution low-voltage side stoichiometric point or high-pressure side stoichiometric point, distribution feeder critical point stoichiometric point record _i, user node reactive power Q _i, user node voltage U _i, feeder line outlet active-power P _gz,, feeder line outlet reactive power Q _gz,, feeder line Egress node voltage U _g;

Step 2, the user node active-power P based on MDMS current time _i, user node reactive power Q _iwith feeder line outlet voltage U _gcarry out trend calculating, obtain the theory wire loss active power Δ P of feeder line _sc, theory wire loss reactive power Δ Q _scwith each node calculating voltage U _{i_c};

Step 3, the user's stealing power calculation based on statistical line losses analytical approach;

Step 4, carry out the judgement of user's electricity filching behavior, if feeder line management line loss active power Δ P _smwith feeder line management line loss reactive power Δ Q _smbeing greater than threshold value δ has stealing to occur, and goes to step 5 and carries out stealing point location, otherwise occur without stealing;

Step 5, node voltage variance analysis;

Step 6, stealing point location and power-steeling quantity calculate, and send stealing warning message.

The active-power P of feeder line outlet for described step 3 _gzdeduct all user's active power of feeder line sum deduct again the feeder line theory wire loss active power Δ P that step 2 calculates _sc, obtain this feeder line management line loss active power Δ P _sm; Reactive power Q with feeder line outlet _gz, deduct all user's reactive powers of feeder line sum deduct again the feeder line theory wire loss reactive power Δ Q that step 2 calculates _sc, obtain this feeder line management line loss reactive power; User's stealing rating formula based on statistical line losses analytical approach is shown below:

$\left\{\begin{array}{c}{\mathrm{\ΔP}}_{\mathrm{sm}}=P_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{P}_i-{\mathrm{\ΔP}}_{\mathrm{sc}}\\ {\mathrm{\ΔQ}}_{\mathrm{sm}}=Q_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{Q}_i-{\mathrm{\ΔQ}}_{\mathrm{sc}}\end{array}\right.$

Wherein, Δ P _smfor feeder line management line loss active power, Δ Q _smfeeder line management line loss reactive power, P _gzfor the active power of feeder line outlet, Q _gzthe reactive power of feeder line outlet.

In described step 4, feeder line management transmission power is greater than actual stealing power and measuring error sum as the higher limit of stealing power, can obtain thus stealing active power codomain [0, Δ P _sm] and the codomain of stealing reactive power [0, Δ Q _sm].

In described step 5 by the calculating voltage U of each load bus calculating by trend _{i_c}with user node voltage U _iit is poor to do, and obtains each node voltage deviation delta U _i, be shown below:

ΔU _i＝U _{i_c}-U _i

Due to the generation of electricity filching behavior, the active power that stealing point is uploaded and reactive power are less than active power and the reactive power of stealing point actual consumption, the U that trend calculates _{i_c}be greater than each user node U _i, and the voltage deviation Δ U of the neighbor node after stealing point and stealing point _ilarger.

In described step 6, be by the Δ U in step 5 _icarry out from big to small monitoring point sequence, adopt pointwise stealing power iteration optimization computing method to carry out stealing point location and power-steeling quantity calculating; Derivation algorithm adopts basic particle group algorithm, and stealing active power and reactive power are variable to be solved, and objective function is each node voltage absolute value of the bias sum of feeder line minimum, hour, power distribution network running status and trend computing mode are on all four, and constraint condition is that stealing active power is less than this feeder line management line loss active power Δ P _sc, stealing reactive power is less than this feeder line management line loss reactive power Δ Q _sm.

The invention has the beneficial effects as follows the data management system of the senior measurement system based on existing power distribution network, electricity consumption data to large user gather computational analysis, calculate feeder line theory wire loss power, node voltage, and judge whether to occur the calculating of stealing and stealing point location and power-steeling quantity, have that measurements and calculations data precision is high, synchronism good, speed is fast, the feature of accurate positioning, can find in real time stealing point and calculate stealing power, accomplish to find at any time to administer at any time, reduce to greatest extent the loss of power supply enterprise.

Accompanying drawing explanation

Fig. 1 is senior measurement architectural schematic;

Fig. 2 is the real-time electricity anti-theft method process flow diagram of the power distribution network large user based on senior measurement system;

Fig. 3 is the senior measurement system arrangement plan of power distribution network;

Fig. 4 is power distribution network topological structure and parameter distribution in embodiment.

Embodiment

Below in conjunction with drawings and Examples, method proposed by the invention is described further.

The invention discloses a kind of real-time electricity anti-theft method of power distribution network large user based on senior measurement system, particular content is as follows:

Target user node voltage U when the intelligent electric meter in the senior measurement system of intelligent distribution network can be the band of each monitoring point (large user's distribution low-voltage side stoichiometric point or high-pressure side stoichiometric point, distribution feeder critical point stoichiometric point) _i, user node active-power P _i, user node reactive power Q _iupload to the MDMS of senior measurement system.

Due to senior measurement system have clock synchronous function (generally by software to time mode realize), target node voltage U during the band that obtains _i, active-power P _i, reactive power Q _iunder the mode of power distribution network steady-state operation, can be used as the data of section at the same time, meet node voltage and Branch Power Flow equation, can carry out distribution power system load flow calculation.

Based on MDMS data constantly, utilize distribution feeder trend computing module to carry out trend calculating, obtain the theory wire loss active power Δ P of feeder line _sc, theory wire loss reactive power Δ Q _scwith each node calculating voltage U _{i_c}.

The active-power P that the critical point intelligent electric meter (as shown in Fig. 3 1.) providing with MDMS is measured _gz, reactive power Q _gzdeduct respectively each user's electric energy meter (as in Fig. 3 2. 3. 4. 5. 6. as shown in) active power sum reactive power sum obtain the statistical line losses active power Δ P of this feeder line _stwith statistical line losses reactive power Δ Q _st, as shown in formula (2).

$\left\{\begin{array}{c}{\mathrm{\ΔP}}_{\mathrm{st}}=P_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{P}_i\\ {\mathrm{\ΔQ}}_{\mathrm{st}}=Q_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{Q}_i\end{array}\right.---\left(2\right)$

Statistical line losses active power Δ P with this feeder line _stwith statistical line losses reactive power Δ Q _stdeduct respectively the theory wire loss active power Δ P of this feeder line _scwith theory wire loss reactive power Δ Q _sc, obtain the management line loss active power Δ P of this feeder line _smwith management line loss reactive power Δ Q _sm, as shown in formula (3).

$\left\{\begin{array}{c}{\mathrm{\ΔP}}_{\mathrm{sm}}={\mathrm{\ΔP}}_{\mathrm{st}}-{\mathrm{\ΔP}}_{\mathrm{sc}}\\ {\mathrm{\ΔQ}}_{\mathrm{sm}}={\mathrm{\ΔQ}}_{\mathrm{st}}-{\mathrm{\ΔQ}}_{\mathrm{sc}}\end{array}\right.---\left(3\right)$

Because the theory wire loss power of feeder line is to calculate according to the measurement performance number of intelligent electric meter, be less than the active loss power (practical line loss power is produced by measurement power and the acting in conjunction of stealing power of electric energy meter) of this feeder line.Therefore, the feeder line that obtained by formula (3) management transmission power is greater than actual stealing power and measuring error sum, can be similar to the higher limit as stealing power, can obtain thus stealing active power codomain [0, Δ P _sm] and the codomain of stealing reactive power [0, Δ Q _sm].

Management line loss equals the difference that statistical line losses deducts theory wire loss, and theory wire loss is inevitably, be that distribution line will consume really, and statistical line losses is the true loss of power supply enterprise.Management transmission power is mainly produced by measuring error and stealing.In intelligence measurement system, the power attenuation that measuring error causes is much smaller than stealing power, and in theory, if do not consider measuring error, while occurring without stealing, the management transmission power of feeder line is zero.Therefore,, if the management transmission power of feeder line is larger, explanation has stealing to occur.

In addition, when stealing electricity phenomenon occurs, (suppose that stealing point occurs in Fig. 3 4. point), 4. put P that intelligent electric meter measures, Q value than actual value (realtime power of user's actual consumption for 4. point measurement power and stealing power sum) less than normal.Due to 4. point measurement to active power and the reactive power active power and the reactive power that are less than user's actual consumption, so the calculating voltage value of each node of calculating of trend can be more higher than the measured value obtaining by electric energy meter.By the calculating voltage U of each load bus calculating by trend _{i_c}with user node U _iit is poor to do, and obtains each node voltage deviation delta U _i, as shown in formula (4).Generation due to electricity filching behavior, the active power that stealing point is uploaded and reactive power are less than active power and the reactive power of stealing point actual consumption, the magnitude of voltage that trend calculates is greater than each monitoring point actual voltage value, and the voltage deviation of the neighbor node after stealing point and stealing point is larger, the possibility of the node generation stealing that voltage deviation is larger is larger.

ΔU _i＝U _{i_c}-U _i (4)

Stealing point location with cut electric power and calculate.With stealing active power and reactive power, for variable to be solved, objective function is each node voltage deviation sum of feeder line minimum hour, power distribution network actual motion status data and the trend computational data based on measurement data are on all four), constraint condition is that stealing active power is less than this feeder line management loss active power Δ P _sc, stealing reactive power is less than this feeder line management loss reactive power Δ Q _sm.Derivation algorithm adopts basic particle group algorithm.

It is below the applied stealing analysis of the method enforcement calculated example of not considering that measuring error adopts the present invention to propose.

The equivalent power distribution network shown in Fig. 4 of take is example, carries out electricity filching behavior judgement and stealing point location, and the branch impedance parameter in figure, transformer impedance parameter are for conversion is on high-tension side parameter.

Reading the measurement data of current time in MDMS comprises: critical point intelligent electric meter measuring voltage amplitude 10.5kV, active power 2134.44kW, reactive power 1086.72kVar (in this simulation example active power 2134.44kW and reactive power 1086.72kVar carry out trend according to user's actual load calculate), the active power of each user's intelligent electric meter, reactive power and node voltage amplitude, as shown in fourth, fifth, six row in table 1.

Active power, the reactive power data that the voltage magnitude of measuring based on critical point intelligent electric meter and each user's intelligent electric meter are measured push back generation calculating before carrying out.Obtain computing node voltage and feeder line theory wire loss power based on measurement data, as shown in table 1 the 7th row and the 8th row.

With active power 2134.44kW, the reactive power 1086.72kVar of critical point intelligent electric meter, deduct respectively feeder line theory wire loss active power 83.74kW, theory wire loss reactive power 128.82kVar, deduct again measurement active power 1830kW and the reactive power 830kVar of each load point, obtain management active loss power 220.70kW and the management reactive loss power 147.90kVar of this feeder line, as shown in table 1 the 9th row.Do not consider in theory measuring error, while occurring without stealing, the management transmission power of feeder line is zero.The management line loss active power that this example obtains is 220.70kW, and management line loss reactive power is 147.90kVar, can conclude and have stealing to occur.

In addition, by calculating each node calculating voltage and uploading the poor of magnitude of voltage, obtain each node voltage deviation, the node that voltage deviation is larger is doubtful stealing node.By last row of table 1, can be found out, 4. 5. 6. 2. 3. the doubtful degree of node stealing node generation stealing is from high to low successively, meets stealing point and the stealing point larger judgment criterion of neighbor node voltage deviation afterwards in step 5.

Table 1 power distribution network stealing data analysis

Stealing point location and stealing power calculation.Suppose that respectively node, 4. 5. 6. for suspicious stealing contact, adopts particle cluster algorithm, with stealing active power and reactive power, for variable to be solved, objective function is each node voltage deviation sum of feeder line minimum, constraint condition is that stealing active power codomain is [0,220.70] kW, stealing reactive power codomain is [0,147.90] kVar.In order to verify that 3. 2. the probability of stealing occur is minimum, to node be 2. 3. also assumed to be stealing point capable power-steeling quantity analysis.The analysis result of each node is as shown in table 2.

Each node voltage deviation sum when table 2 stealing occurs in different node

4. the load point of voltage deviation absolute value sum minimum is stealing point, while supposing 2. 3. for stealing node, $\underset{i=1}{\overset{N}{\mathrm{\Σ}}}\left|{\mathrm{\ΔU}}_i\right|$ Larger.

The error analysis of the inventive method:

The middle-and-high-ranking measurement system of practical application has measuring error, and the error of mark is electric energy metering error conventionally, and the precision of senior measurement system is all in 0.5 grade, and the precision that special gate energy meter has reaches 0.2 grade or 0.1 grade.The measuring accuracy of the power of senior measurement system, voltage, electric current will be far above electric energy metrical precision, conventionally in 0.2 grade.

Therefore the stealing analysis of considering measuring error is the key point of weighing the method practicality.

The measurement relative error of supposing power and voltage is below respectively ± 0.2% and ± 0.1% carry out simulation calculation, and result is respectively if table 3 is to as shown in table 6.

Simulation analysis result when table 3 measurement relative error is ± 0.2%

Each node voltage deviation sum when table 4 stealing occurs in different node

Simulation analysis result when table 5 measurement relative error is ± 0.1%

Each node voltage deviation sum when table 6 stealing occurs in different node

By table 4 and table 6, can be found out, the senior measurement system based on 0.2 grade and 0.1 grade, can accurately analyze stealing position and stealing power.

In addition, the method computational accuracy is also relevant with user's power-steeling quantity size, through simulation calculation, power-steeling quantity (applied power) is greater than 50kVA, and when measuring accuracy is 0.2 grade, system can accurately be judged stealing point and stealing power, when measuring accuracy is 0.1 grade, the stealing that is greater than 20kVA can accurately be judged.When not meeting above-mentioned requirements, stealing is analyzed accurate qualitative meeting and is declined to some extent.While being 0.2 grade such as the precision of the measuring system adopting in above-mentioned example, when power-steeling quantity power is less than 50kVA, the stealing point of judging may be sometimes node 5..

About the stability analysis of particle swarm optimization algorithm result of calculation.With particle cluster algorithm, carry out multivariate optimization problem and release, exist each result of calculation inconsistent, this is because the randomness of particle cluster algorithm produces, that is, differ and find surely optimum solution, and be normal phenomenon.When optimization problem is relatively simple, each result of calculation can be consistent,, can find optimum solution that is at every turn.Variable to be optimized of the present invention only has two, and value constraint within the specific limits.Actual analysis finds, population scale is 60, and when maximum iteration time is 30 times, result of calculation is unique, while finding the mean iterative number of time of optimum solution 10 times.

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

Claims (5)

1. the real-time electricity anti-theft method of power distribution network large user based on senior measurement system, is characterized in that, mainly comprises:

Step 1, read the metric data of each intelligent electric meter of MDMS current time, mainly comprise: the user node active-power P that each large user's distribution low-voltage side stoichiometric point or high-pressure side stoichiometric point, distribution feeder critical point stoichiometric point record _i, user node reactive power Q _i, user node voltage U _i, feeder line outlet active-power P _gz, feeder line outlet reactive power Q _gz, feeder line Egress node voltage U _g;

Step 2, the user node active-power P based on MDMS current time _i, user node reactive power Q _iwith feeder line Egress node voltage U _gcarry out trend calculating, obtain the theory wire loss active power Δ P of feeder line _sc, theory wire loss reactive power Δ Q _scwith each node calculating voltage U _{i_c};

Step 3, the user's stealing power calculation based on statistical line losses analytical approach;

Step 4, carry out the judgement of user's electricity filching behavior, if feeder line management line loss active power Δ P _smwith feeder line management line loss reactive power Δ Q _smbeing greater than threshold value δ has stealing to occur, and goes to step 5 and carries out stealing point location, otherwise occur without stealing;

Step 5, node voltage variance analysis;

Step 6, stealing point location and power-steeling quantity calculate, and send stealing warning message.

2. method according to claim 1, is characterized in that, the active-power P of feeder line outlet for described step 3 _gzdeduct all user's active power of feeder line sum deduct again the feeder line theory wire loss active power Δ P that step 2 calculates _sc, obtain this feeder line management line loss active power Δ P _sm; Reactive power Q with feeder line outlet _gz, deduct all user's reactive powers of feeder line sum deduct again the feeder line theory wire loss reactive power Δ Q that step 2 calculates _sc, obtain this feeder line management line loss reactive power; User's stealing rating formula based on statistical line losses analytical approach is shown below:

$\left\{\begin{array}{c}{\mathrm{\ΔP}}_{\mathrm{sm}}=P_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{P}_i-{\mathrm{\ΔP}}_{\mathrm{sc}}\\ {\mathrm{\ΔQ}}_{\mathrm{sm}}=Q_{\mathrm{gz}}-\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{Q}_i-{\mathrm{\ΔQ}}_{\mathrm{sc}}\end{array}\right.$

Wherein, Δ P _smfor feeder line management line loss active power, Δ Q _smfeeder line management line loss reactive power, P _gzfor the active power of feeder line outlet, Q _gzthe reactive power of feeder line outlet.

3. method according to claim 1, is characterized in that, in described step 4, feeder line management transmission power is greater than actual stealing power and measuring error sum as the higher limit of stealing power, can obtain thus stealing active power codomain [0, Δ P _sm] and the codomain of stealing reactive power [0, Δ Q _sm], wherein, Δ P _smfor feeder line management line loss active power, Δ Q _smfeeder line management line loss reactive power.

4. method according to claim 1, is characterized in that, in described step 5 by the calculating voltage U of each load bus calculating by trend _{i_c}with user node voltage U _iit is poor to do, and obtains each node voltage deviation delta U _i, be shown below:

ΔU _i＝U _{i_c}-U _i

Due to the generation of electricity filching behavior, the active power that stealing point is uploaded and reactive power are less than active power and the reactive power of stealing point actual consumption, the U that trend calculates _{i_c}be greater than each user node voltage U _i, and the voltage deviation Δ U of the neighbor node after stealing point and stealing point _ilarger.

5. method according to claim 1, is characterized in that, is by each node voltage deviation delta U in step 5 in described step 6 _icarry out from big to small monitoring point sequence, adopt pointwise stealing power iteration optimization computing method to carry out stealing point location and power-steeling quantity calculating; Derivation algorithm adopts basic particle group algorithm, and stealing active power and reactive power are variable to be solved, and objective function is each node voltage absolute value of the bias sum of feeder line minimum, hour, power distribution network running status and trend computing mode are on all four, and constraint condition is that stealing active power is less than this feeder line management line loss active power Δ P _sc, stealing reactive power is less than this feeder line management line loss reactive power Δ Q _sm.