CN106655152A - Power distribution network state estimation method based on AMI measurement characteristics - Google Patents

Abstract

The invention discloses a power distribution network state estimation method based on AMI measurement characteristics, comprising the following steps: (1) determining a system model and parameters thereof according to the structure of a power distribution network, and simulating SCADA measurement and AMI measurement in the system through use of a typical daily load curve and a true value of power flow according to the characteristics of different measurement; (3) determining the cycle of power distribution network state estimation under the condition that SCADA measurement data and AMI measurement data match each other; and (4) on the basis of the measurement data determined in steps (2) and (3), estimating the state of the power distribution network, and determining the running state of the system. By fully considering the measurement characteristics of AMI, the invention presents a solution to the problem that AMI measurement data delay and SCADA data measurement cycle are inconsistent in state estimation. Power distribution network state estimation through comprehensive use of SCADA and AMI measurement data is realized.

Description

A distribution network state estimation method based on AMI measurement characteristics

technical field

The invention belongs to the field of power system analysis, in particular to a method for estimating the state of a power distribution network based on AMI measurement characteristics.

Background technique

State estimation is a core function of energy management systems. According to the application scenario, state estimation can be divided into two modes: online and offline. The time interval of distribution network online state estimation depends on the power company, generally 10 to 60 minutes, and is mainly used for distribution network scheduling and energy management to support real-time decision-making and applications. Its characteristic is to use as accurate and real-time data as possible to analyze the current situation of the system; offline state estimation is to analyze the situation of the power grid in a certain period of time in the past based on all the measurements obtained, and is mainly used for power grid analysis, power theft analysis, etc.

In the field of distribution network state estimation, a large number of scholars have carried out extensive research. According to different state estimation models, it can be roughly divided into state estimation algorithms with node voltage, branch current, and branch power as state variables.

Due to the lack of measurement data of the low-voltage distribution network on the user side, in the state estimation of the distribution network, the state estimation is usually only performed on the medium-voltage distribution network. Since 2009, the State Grid Corporation of China has taken "full coverage, full collection, and full cost control" as its construction goal to promote the application of smart energy meters and the construction of electricity information collection systems. As of the end of October 2013, a total of 173 million smart energy meters have been installed and applied, and the electricity consumption information collection system has covered 173 million households. The promotion of smart meters provides a large number of redundant measurements for low-voltage distribution networks. This information enriches the measurement types of the distribution network, and can effectively solve the long-standing problem of unobservable problems of a large number of feeders and their branches caused by insufficient configuration of measurement devices and imperfect communication channels. Reasonable application of the data collected by AMI to distribution network state estimation can estimate various measurement and state information more accurately and comprehensively, thus providing complete, reliable and high-precision analysis for higher-level applications data.

However, compared with SCADA measurements, AMI data has its unique measurement characteristics:

1) The measurement interval of power distribution SCADA measurement is generally within 20s, a few minutes at most; the interval of AMI measurement can be preset, generally 15min, 30min or one hour.

2) The smart meter has two reading methods. Most foreign countries use the freezing method, that is, pre-set the freezing time, and then read it back. The data in each table is time-stamped, but the read-back time is uncertain. The time delay of reading data from each table at the same time is within 20s or even lower. In our country, only the data at zero o'clock every day is frozen, and the rest of the time is recruited and read, that is, the metering center sends instructions to read the meters in turn, and read one meter before reading the other. It takes 10 minutes to 15 minutes to read the electric meter in one station area according to the number of users.

3) The accuracy level of SCADA measurement is generally about 2; the level of smart meters used for AMI measurement is generally 0.5 or even higher.

In practice, the measurement of AMI has problems such as inconsistency of data time scale or time delay, and inconsistency with the measurement cycle of SCADA data. These issues are the key issues that limit the application of AMI data to distribution network state estimation.

The present invention considers the application practice of state estimation in China and the reading mode of AMI, and focuses on analyzing the off-line state estimation using recruited AMI data.

Contents of the invention

At present, scholars have conducted research on the application of AMI measurement data in distribution network state estimation. However, the existing studies failed to fully consider the measurement characteristics of AMI. The present invention starts from the requirements of distribution network state estimation and the actual characteristics of AMI measurement data, and aims at the offline state estimation application scenario of applying to read AMI data, and proposes AMI measurement data delay and SCADA data measurement cycle in state estimation The solution to the inconsistency and other problems realizes the state estimation of distribution network by comprehensively utilizing SCADA and AMI measurement data. Combined with the status quo of domestic AMI measurement reading, the error situation of distribution network state estimation before and after AMI delayed data processing is analyzed; the influence of adjustment of state estimation cycle on the result error is analyzed; the influence of measurement noise on state estimation results is analyzed.

In order to solve the above technical problems, the present invention provides a distribution network state estimation method based on AMI measurement characteristics, comprising the following steps:

Step 1. The distribution network state estimation scenario is to apply the offline state estimation of the AMI data, determine the system model and parameters according to the distribution network structure, and use the typical daily load curve and the true value of the power flow to simulate the SCADA measurement and AMI in the system The maximum noise of SCADA measurement is 2%, the maximum noise of AMI measurement is 0.5%, the time interval of SCADA measurement is set to 1min, and the time interval of AMI measurement is 15min;

Step 2, delay processing the AMI measurement data;

In the offline state estimation scenario where AMI data is recruited and read, the system has collected electric energy measurement data before and after each time point. For active power, the average active power calculated from the electric energy value of the smart meter is used instead of the instantaneous active power.

In formula (1), for each smart meter, t ₀ and t ₁ are its adjacent measurement and reading moments; P is the instantaneous active power at each moment from t ₀ to t ₁ ; W is the two values of t ₀ and t ₁ The difference value measured by the electric energy of the electric meter at a time; is the average active power from t ₀ to t ₁ ;

to average active power The data is corrected, the network loss is ignored, and the AMI average active power of each user in the same station area on the low-voltage side is superimposed from bottom to top to obtain the superimposed average active power of the station substation; at the same time, a real-time SCADA is installed at the station substation Measurement, for the same point at the same time, get a real-time SCADA measurement _PS and _a superimposed active power PA; use the real _- time measurement _PS to correct the superimposed active power PA, that is, multiply the superimposed active power of the substation by a Correction factor Correspondingly, the average active power allocated to each user is also multiplied by the correction factor;

The power factor of the user is 0.95-0.98, and the power factor of the user remains unchanged during the time period from t ₀ to t _1. According to the average active power of the user and power factor to calculate the average reactive power of the user

Step 3. Determine the cycle of distribution network state estimation when the SCADA measurement and AMI measurement data cooperate, that is, use the SCADA measurement at each moment to predict the AMI measurement value at the same moment;

The period of SCADA measurement is T _s , the period of AMI measurement is T _A , and T _A is greater than T _s ;

(1) At time T ₀ , the user's AMI electric energy measurement data is obtained, and the average active power of each user is calculated by the method of step 2 and average reactive power

(2) During the time period from T ₀ to T ₀ +T _A , for the nth SCADA measurement time T ₀ +nT _s , obtain the SCADA active power measurement value, according to the average active power among different users at T ₀ Accounting for the percentage of SCADA active power measurement P _S at this time, the SCADA active power measurement value at T ₀ +nT _s is distributed to each user, and the AMI active power prediction value of each user at T ₀ +nT _s is obtained; and then through the user The power factor is calculated to obtain the AMI reactive power prediction value of each user at T ₀ +nT _s time;

Step 4: Estimate the state of the distribution network to determine the operating state of the system, use the SCADA measurement data and AMI measurement data processed in steps 2 and 3 as input, estimate the near state of the distribution network, and determine the operation of the system state.

Compared with prior art, the beneficial effect of the present invention is:

In practice, the measurement of AMI has problems such as inconsistency of data time scale or time delay, and inconsistency with the measurement cycle of SCADA data. These issues are the key issues that limit the application of AMI data to distribution network state estimation. Compared with other scholars' research on AMI data in the field of distribution network state estimation, the present invention has the following advantages: fully consider the measurement characteristics of AMI, mainly including AMI measurement delay and cooperation with SCADA measurement, and propose corresponding A method for data processing and determining the state estimation period.

Description of drawings

Fig. 1 is 13 node system wiring diagrams provided by the present invention;

Fig. 2 is the relative error of state estimation voltage amplitude at time 1 before and after AMI data delay processing provided by the present invention;

Fig. 3 is the state estimation voltage phase angle absolute error of time 1 before and after AMI data delay processing provided by the present invention;

Fig. 4 is the estimated value of voltage at point n1 and the actual voltage measurement value of SCADA at point m7 provided by the present invention.

detailed description

The technical solution of the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the described specific embodiments are only for explaining the present invention, and are not intended to limit the present invention.

A kind of distribution network state estimation method based on AMI measurement characteristic of the present invention comprises the following steps:

Step 1. The distribution network state estimation scenario is to apply the offline state estimation of the AMI data, determine the system model and parameters according to the distribution network structure, and use the typical daily load curve and the true value of the power flow to simulate the system in accordance with the characteristics of different measurements. SCADA measurement and AMI measurement, wherein, the maximum noise of SCADA measurement is 2%, the maximum noise of AMI measurement is 0.5%, the time interval of SCADA measurement is set to 1min, and the time interval of AMI measurement is 15min .

Step 2. Delay processing of AMI measurement data; in the offline state estimation scenario where the application recruits and reads AMI data, the system has collected electric energy measurement data before and after each time point. For active power, the electric energy value of the smart meter is calculated The resulting average active power replaces the instantaneous active power,

In formula (1), for each smart meter, t ₀ and t ₁ are its adjacent measurement and reading moments; P is the instantaneous active power at each moment from t ₀ to t ₁ ; W is the two values of t ₀ and t ₁ The difference value measured by the electric energy of the electric meter at a time; is the average active power from t ₀ to t ₁ .

to average active power The data is corrected, the network loss is ignored, and the AMI average active power of each user in the same station area on the low-voltage side is superimposed from bottom to top to obtain the superimposed average active power of the station substation; at the same time, a real-time SCADA is installed at the station substation Measurement, for the same point at the same time, get a real-time SCADA measurement _PS and _a superimposed active power PA; use the real _- time measurement _PS to correct the superimposed active power PA, that is, multiply the superimposed active power of the substation by a Correction factor Correspondingly, the average active power allocated to each user is also multiplied by the correction factor.

According to historical statistical data, the power factor of the user is 0.95-0.98, it is considered that the power factor of the user is constant in the time period from t ₀ to t ₁ , and the present invention uses the average active power of the user obtained and power factor to calculate the average reactive power of the user

Step 3. Determine the cycle of distribution network state estimation when the SCADA measurement and AMI measurement data are coordinated, that is, use the SCADA measurement at each moment to predict the AMI measurement value at the same moment; the SCADA measurement cycle is T _s , the period of AMI measurement is T _A , in practice T _A is greater than T _s .

(1) At time T ₀ , the user's AMI electric energy measurement data is obtained, and the average active power of each user is calculated by the method of step 2 and average reactive power

(2) During the time period from T ₀ to T ₀ +T _A , for the nth SCADA measurement time T ₀ +nT _s , obtain the SCADA active power measurement value, according to the average active power among different users at T ₀ Accounting for the percentage of SCADA active power measurement P _S at this time, the SCADA active power measurement value at T ₀ +nT _s is distributed to each user, and the AMI active power prediction value of each user at T ₀ +nT _s is obtained; and then through the user The power factor is calculated to obtain the AMI reactive power prediction value of each user at T ₀ +nT _s time;

Step 4. Estimating the state of the distribution network to determine the operating state of the system, including: taking the SCADA measurement data and AMI measurement data processed in steps 2 and 3 as input, estimating the near state of the distribution network, and determining the system operating status.

research material:

Taking the IEEE13 node system as an analysis example, as shown in Figure 1. m2 to m7 are the medium-voltage distribution network, each node is set for SCADA measurement, n1 to n6 is the user side, and each node is set for AMI measurement.

The calculation example includes the accurate measurement data of each point and the true value of the power flow at a total of 15 moments from 1 to 15 minutes. Add random noise obeying Gaussian distribution to each measurement to simulate measurement error, where the maximum noise of SCADA measurement is 2%, and the maximum noise of AMI measurement is 0.5%. Set the time interval of SCADA measurement as 1min, and the time interval of AMI measurement as 15min. In order to simulate the influence of time delay, the AMI data of different nodes obtained by ordering recruitment at time 1 are all under different time scales. The arrival time of the AMI data of each node differs by 1 minute.

Use the method in step 2 to process the AMI delay data at time 1, and then perform state estimation on the medium and low voltage hybrid network. Figure 2 and Figure 3 show the error situation between state estimation and power flow true value before and after delay processing. The result shows that, because the AMI measurement is delayed, the error of the state estimation will be greatly increased by directly using the delay data for state estimation; after the delay processing method proposed by the present invention is used to process the data, the error of the state estimation is reduced to an acceptable level within range.

Use the method in step 3 to determine the calculation period of the state estimation, that is, use the SCADA measurements at 15 moments to estimate the AMI measurements at 15 moments. Taking node n1 as an example, the comparison between its estimated voltage value and the actual voltage measured by SCADA at point m7 at different times is shown in Figure 4. Wherein, the dotted line part is a schematic curve of voltage change after 15 minutes. The state estimation of the distribution network is carried out on the calculation example within 15 time periods. The relative error percentages between the voltage amplitude of the nodes and the true value of the power flow at different times, and the absolute error between the voltage phase angle and the true value of the power flow are shown in Table 1 and Table 1. 2.

Table 1 The relative error percentage of estimated voltage amplitude of each node state at different times

Table 2 Absolute error percentage of voltage phase angle estimation for each node state at different times

It can be seen from Table 1 and Table 2 that by using the above method to estimate the AMI measurement value and then perform state estimation, the state results at multiple times can be obtained, and the change of the system state can be described in more detail. Since nodes n1-n6 are measured by AMI, their power values are all estimated values, so the error is larger than that of nodes m2-m7, but it is still within an acceptable range. It should be noted that the present invention describes a simple method for pseudo-measurement modeling with the help of AMI data. If the accuracy of load forecasting is improved, the calculation results of the method of the present invention will be more accurate, and the operating state of the power system will be more fully obtained. accurate description.

In order to further verify the effectiveness of the present invention, 100 groups of data meeting the above-mentioned measurement accuracy were randomly simulated. Each set of data includes various types of measurements of 13 nodes in the example within 15 minutes, and the measurement noise of different nodes is different. Due to the large amount of data, the present invention takes the n1 node with a relatively large voltage amplitude error as an example to describe the influence of noise conditions on the state estimation accuracy. Table 3 reflects the change of the percentage of the relative error of the n1 node voltage amplitude under 100 measurement noise scenarios. On the whole, the fluctuation range of the relative error of the voltage amplitude caused by the noise change is between 0.10% and 0.20%.

Table 3 Relative error of voltage amplitude at point n1 under 100 sets of measurement noise combinations

It can be seen from Table 3 that the voltage amplitude error from time 3 to time 6 is generally small. During this period of time, the node load changes slowly, and the error caused by the AMI measurement and prediction method of step three is small, and the influence on the accuracy of the state estimation is also small.

Although the present invention has been described above in conjunction with the accompanying drawings, the present invention is not limited to the above-mentioned specific embodiments, and the above-mentioned specific embodiments are only illustrative, rather than restrictive. Under the enlightenment of the present invention, many modifications can be made without departing from the gist of the present invention, and these all belong to the protection of the present invention.

Claims (1)

1. it is a kind of based on AMI measure characteristic State Estimation for Distribution Network, it is characterised in that：Comprise the following steps：

Step one, state of electric distribution network estimate that scene is to estimate using the off-line state for recruiting reading AMI data, true according to distribution net work structure Determine system model, parameter, measured using the SCADA in typical day load curve and trend true-value simulation system and AMI is measured, its In, it is 0.5% that the maximum noise that SCADA is measured is the maximum noise that 2%, AMI is measured, and sets the time interval that SCADA is measured For 1min, the time interval that AMI is measured is 15min；

Step 2, delay disposal is carried out to AMI metric data；

Recruit in application and read under the off-line state estimation scene of AMI data, system has collected the electric energy before and after each time point and measured Data, for active power, by the calculated average active power of the energy value of intelligent electric meter instantaneous active power are replaced,

$\stackrel{\‾}{P}=\frac{{\∫}_{t_0}^{t_1}Pdt}{t_1-t_0}=\frac{W}{t_1-t_0}---\left(1\right)$

In formula (1), to each intelligent electric meter, t₀、t₁The moment is read for its adjacent measurement；P is t₀To t₁Each moment instantaneously has Work(power；W is t₀、t₁The difference that two moment ammeter electric energy are measured；For t₀To t₁Average active power；

To average active powerData are modified, and ignore network loss and disregard, and the AMI of each user in same area of low-pressure side is average Active power stacks up from bottom to up, obtains the superposed average active power of platform change；Meanwhile, it is also equipped with platform change in real time SCADA measure, for the same point of synchronization, obtain a real-time SCADA measurement P_SWith a superposition active-power P_A； Use real-time measurement P_SAmendment superposition active-power P_A, i.e., a correction factor is multiplied by the superposition active power that platform becomesPhase Answer, the average active power for being assigned to each user is also multiplied by the correction factor；

The power factor value of user is 0.95-0.98, and in t₀To t₁The power factor of user is constant in time period, according to Family average active powerWith the average reactive power that power factor calculates user

Step 3, determine the cycle that state of electric distribution network is estimated in the case that SCADA is measured and AMI metric data coordinates, i.e., with every The SCADA at individual moment measures prediction AMI measuring values in the same time；

The cycle that SCADA is measured is T_s, the cycle that AMI is measured is T_A, T_AMore than T_s；

(1) in T₀At the moment, the AMI electric energy metric data of user is obtained, the average of each user is calculated by the method for step 2 Active powerAnd average reactive power

(2) in T₀To T₀+T_AIn time period, for n-th SCADA measures moment T₀+nT_s, obtain the active measuring values of SCADA, root According to T₀Average active power between moment different userAccount for the active measurement P of moment SCADA_SPercentage, by T₀+nT_sMoment The active measuring values of SCADA distribute to each user, try to achieve T₀+nT_sThe AMI active power predicted values of each user at moment；Again by using The power factor at family, is calculated T₀+nT_sThe AMI reactive power predicted values of each user at moment；

Step 4, state estimation is carried out to power distribution network, determine the running status of system, after being processed with step 2 and step 3 SCADA metric data and AMI metric data, to the nearly state estimation of power distribution network, determine the running status of system as input.