A distribution network status assessment method based on multi-section intelligent instrument data

Abstract

A distribution network status evaluation method based on multi-section intelligent instrument data, characterized in that the method comprises the following steps: Step 1, obtaining the AMI sampling time, and determining the reading of SCADA sampling data and AMI sampling data based on the comparison between the current time and the AMI sampling time; wherein, when the AMI sampling data cannot be read, the AMI pseudo-measurement data is fitted by a polynomial fitting method based on the preset sliding time window length and polynomial fitting order; Step 2, based on the data obtained in step 1, the real-time status of the system node of the distribution network is evaluated, and the evaluation result is output. The method of the present invention has a simple operation process, the pseudo-measurement result can be flexibly and smoothly fine-tuned based on historical data, and the evaluation result is accurate.

Description

Technical Field

[0001] The present invention relates to the field of power system control, and more specifically, to a distribution network state evaluation method based on multi-section intelligent meter data. Background Art

[0002] At present, in the context of promoting the transformation of energy structure, a large number of distributed power sources, electric vehicle charging piles, and energy storage systems are connected to the distribution network in a distributed manner, realizing the input and output of electric energy in a flexible and decentralized manner. In order to meet the power transmission and distribution needs of more complex distribution network systems, it is crucial to obtain the real-time status of the distribution network in a timely and accurate manner.

[0003] In the existing technology, Supervisory Control And Data Acquisition (SCADA) has a wide range of applications, mature technology, and occupies an important position in the application of power systems. As a real-time data source for power system automation, the SCADA system can integrate the control, signal, measurement, billing and other circuits of the substation into the computer system with RTU (Remote Terminal Unit) and microcomputer protection devices as the core, thereby improving the reliability of the secondary system.

[0004] In addition, Advanced Metering Infrastructure (AMI) is an automatic two-way communication architecture between smart meters with IP addresses and power companies. AMI can provide real-time energy consumption data and allow users to choose energy use based on price during use. AMI can be implemented using various technologies such as low-cost radio, RF (Radio Frequency) technology and PLC (Power Line Communication).

[0005] However, there are obvious disadvantages when using SCADA and AMA separately or as a combination of distribution network measurement systems.

[0006] First, in medium and low voltage distribution networks, the number of various measurement devices installed is limited, and the redundancy of measurement data is low. It is difficult to make the evaluation results of the distribution network status meet the requirements of observability by using any single metering architecture.

[0007] Secondly, when two different metering architectures are used in combination, although the data collected by the two types of equipment can be used simultaneously, the sampling frequencies of the two metering architectures are different and vary greatly. Therefore, in order to use the two types of data simultaneously, the time interval for distribution network status evaluation must be greatly extended or the data must be preprocessed to meet the time interval requirements for evaluation. Generally speaking, the sampling frequency of the SCADA system is around 1Hz, while the sampling time interval of AMI is around 15min. If the data is not preprocessed, the evaluation time interval needs to be extended to 15min, which reduces the real-time performance of the distribution network status evaluation.

[0008] Third, although there are methods to generate AMI pseudo-measurement data during non-AMI sampling periods, most studies usually use the AMI sampling value closest in time as the pseudo-measurement data for the subsequent non-AMI sampling period, or generate pseudo-measurement data through simple processing. The pseudo-measurement data generated by this processing method has very low accuracy and cannot meet the requirements for high-precision status assessment of the distribution network during the non-AMI sampling period.

[0009] In view of the above content, the present invention provides a distribution network status assessment method based on multi-section intelligent instrument data. Summary of the invention

[0010] In order to solve the deficiencies in the prior art, the object of the present invention is to provide a distribution network status assessment method based on multi-section intelligent meter data.

[0011] The present invention adopts the following technical solution.

[0012] A distribution network status assessment method based on multi-section intelligent instrument data, wherein the method comprises the following steps: step 1, obtaining the AMI sampling time, and determining the reading of SCADA sampling data and AMI sampling data based on the comparison between the current time and the AMI sampling time; when the AMI sampling data cannot be read, the AMI pseudo-measurement data is fitted by a polynomial fitting method based on a preset sliding time window length and a polynomial fitting order; step 2, based on the data obtained in step 1, the real-time status of the system nodes of the distribution network is assessed, and the assessment result is output.

[0013] Preferably, in step 1, the current time is (t, t+T, t+2T, ..., t+nT), respectively; wherein T is the sampling period for sampling AMI and SCADA data in the distribution network state assessment method, t is the initial assessment time, and n is any natural number.

[0014] Preferably, T is an integer multiple of the sampling interval of the SCADA sampling data, and T is much smaller than the sampling interval of the AMI sampling data.

[0015] Preferably, evaluating the real-time status of the system nodes of the distribution network also includes acquiring AMI and SCADA data at each current moment, and acquiring an evaluation result of the distribution network status based on the acquired data; the AMI data includes AMI sampling data and AMI pseudo-measurement data.

[0016] Preferably, if the current moment is an AMI sampling moment, then AMI sampling data and SCADA sampling data are obtained; if the current moment is not an AMI sampling moment, then AMI pseudo-measurement data and SCADA sampling data are obtained.

[0017] Preferably, before determining the reading of SCADA sampling data and AMI sampling data, the method reads the initial network parameters of the distribution network and establishes the distribution network topology map based on the initial network parameters; the initial network parameters include the line connection relationship of the distribution network, the installation location of the SCADA equipment, and the installation location of the AMI equipment.

[0018] Preferably, the SCADA sampled data includes branch active power and branch reactive power; the AMI sampled data and the AMI pseudo-measured data include node injected active power and node injected reactive power.

[0019] Preferably, the sliding time window length and the polynomial fitting order are updated at each AMI sampling moment; wherein, the updating rule steps of the sliding time window length include: step 1.1, performing one-bit increase or decrease operations on the sliding time window length and the polynomial fitting order of the previous AMI sampling moment, and performing permutations and combinations on the operation results; step 1.2, using all possible results in the permutations and combinations generated in step 1.1 to fit the historical AMI sampling data, and respectively obtaining the errors between the fitting values ​​of different possible results and the sampling values ​​of the AMI sampling data at the current moment; step 1.3, selecting the possible result corresponding to the minimum error, and using the sliding time window length and the polynomial fitting order in the possible result as the fitting parameters of the AMI pseudo-measurement data between the current AMI sampling moment and the next AMI sampling moment.

[0020] Preferably, the fitting parameters of the AMI pseudo-measurement data between the current AMI sampling moment and the next AMI sampling moment obtained in step 1.3 are used to generate corresponding AMI pseudo-measurement data at each current moment to achieve status evaluation of the distribution network.

[0021] Preferably, the fitting function is solved using the least squares method or the weighted least squares method.

[0022] Preferably, when the weighted least square method is used, the node voltage complex phasor is used as the state variable x=[x1, x2, ..., x n ]T , using the measurement data as the measurement vector z = [z1, z2, ..., z m ] T , where z = h(x) + v, the calculation meets the objective function minJ(x) = min[zh(x)] T R -1 [zh(x)] is the value of x under the condition; where h(x) is the nonlinear measurement function of the state vector, v is Gaussian white noise, R -1 is the measurement weight matrix.

[0023] Preferably, initialization is performed before obtaining the AMI pseudo-measurement data, and initial values ​​of the sliding time window length and the polynomial fitting order are set during the initialization process.

[0024] Preferably, when the weighted least squares method is adopted, the initial values ​​of the state variables and the convergence accuracy of the state evaluation are set during the initialization process.

[0025] Preferably, the initial value s1 of the sliding time window length ranges from [2, 10], the initial value p1 of the polynomial fitting order ranges from [1, 5]; and the parameters satisfy the condition s1≥p1+1.

[0026] Preferably, each window in the sliding time window contains a historical AMI sampling data or a current AMI sampling data; the historical AMI sampling data will continuously change with the addition of new AMI sampling data and the length of the sliding time window.

[0027] Preferably, the mean absolute error is used and The accuracy of the pseudo-measurement data is checked; where N is the number of system nodes in the distribution network, M is the number of state evaluations, which depends on the end time and period T of the state evaluation set during the initialization process, and P is the active power. Pseudo measurement data The active power measurement value corresponding to the i-th state evaluation node at the j-th evaluation; Q is the reactive power, Pseudo measurement data The reactive power measurement value corresponding to the i-th state evaluation node at the j-th evaluation.

[0028] Preferably, when the sampling period of SCADA data is 1 s and the sampling period of AMI data is 15 min, the sampling period T for sampling AMI and SCADA data is set to 1 min.

[0029] Preferably, the mean absolute error is used to check the accuracy of the pseudo-measurement data; the maximum absolute error of the voltage amplitude, the maximum absolute error of the voltage phase angle, the mean absolute error of the voltage amplitude and the mean absolute error of the voltage phase angle of the pseudo-measurement data are used to check the accuracy of the evaluation result.

[0030] The beneficial effect of the present invention is that, compared with the prior art, the distribution network status method based on multi-section intelligent instrument data in the present invention collects AMI sampling data and SCADA sampling data to evaluate the system node status of the distribution network in real time, and extracts the most suitable AMI pseudo-measurement data by dynamically adjusting the sliding window length and the polynomial order, thereby realizing high-precision distribution network status evaluation. The method of the present invention has a simple calculation process, the pseudo-measurement results can be flexibly and smoothly fine-tuned based on historical data, and the evaluation results are accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A schematic diagram of the steps in a distribution network status assessment method based on multi-section intelligent instrument data of the present invention;

[0032] Figure 2 A topological diagram of the installation locations of SCADA equipment and AMI equipment and the network architecture of the distribution network in a distribution network status assessment method based on multi-section intelligent instrument data of the present invention;

[0033] Figure 3 It is a schematic diagram of updating a sliding time window in a distribution network state assessment method based on multi-section intelligent instrument data according to the present invention;

[0034] Figure 4 It is a schematic diagram of the changes of two parameters, the basic window number and the fitting order, over time in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention;

[0035] Figure 5 It is a schematic diagram of comparison between the AMI pseudo-measurement data obtained in the second mode of the distribution network state assessment method based on multi-section intelligent instrument data of the present invention and the real injected power of the distribution network;

[0036] Figure 6 It is a schematic diagram of comparison between the AMI pseudo-measurement data obtained in the third mode of the distribution network state assessment method based on multi-section intelligent instrument data of the present invention and the real injected power of the distribution network;

[0037] Figure 7 A schematic diagram for comparing average errors of node three-phase voltage amplitudes in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention;

[0038] Figure 8It is a schematic diagram for comparing the average errors of the phase angles of the three-phase voltages at nodes in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention. DETAILED DESCRIPTION

[0039] The present application is further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present application.

[0040] Figure 1 The figure is a schematic diagram of the steps in a distribution network status assessment method based on multi-section intelligent instrument data of the present invention. Figure 1 As shown, a distribution network status assessment method based on multi-section smart meter data includes step 1 and step 2.

[0041] Step 1, obtain the AMI sampling time, and determine the reading of SCADA sampling data and AMI sampling data based on the comparison between the current time and the AMI sampling time. When the AMI sampling data cannot be read, the AMI pseudo-measurement data is fitted by a polynomial fitting method based on the preset sliding time window length and polynomial fitting order.

[0042] It can be understood that the method of the present invention is similar to the common method of aggregating measurement data of multiple systems for distribution network status assessment often used in the prior art, and both methods can collect data based on a preset status assessment time interval.

[0043] In one embodiment of the present invention, the time interval for evaluating the state of the distribution network can be consistent with the time interval for sampling data of SCADA. By configuring the same time interval, the method of the present invention can accurately obtain all data collected by the SCADA system. Since the sampling interval of the AMI system is generally much larger than the sampling interval of the SCADA system, the interval for evaluating the state of the distribution network in the present invention is based on an integer multiple of the sampling interval of the SCADA system.

[0044] Since the sampling interval of the AMI system data is relatively large, it is not necessarily possible to fully obtain the AMI measurement data at each sampling point of the distribution network status assessment. The present invention adopts a polynomial fitting method to achieve the acquisition of AMI pseudo-measurement data. Different from the method in the prior art that simply uses the AMI measurement data collected in the historical period directly or after simple processing as the pseudo-measurement data of the next period, the present invention performs polynomial fitting of the data by dynamically adjusting the sliding time window length and the fitting order.

[0045] The method of the present invention adopts a method of dynamically adjusting the sliding window length and the fitting order, and its effect is better than the evaluation method set to a fixed length and a fixed order. The parameters adjusted in the method of the present invention include two parameters, the sliding window length and the fitting order, which are determined according to the results of the pre-adjustment, and the two parameters are adjusted at the same time. At the AMI sampling moment, the sliding time window length and the fitting order used in the previous non-AMI sampling time period are fine-tuned to form a variety of combinations, and data fitting is performed on multiple AMI historical sampling values. The current AMI sampling value is used as the true value criterion, and multiple fitting errors at the current moment are calculated, and the sliding time window length and the fitting order corresponding to the minimum fitting error are used as new fitting parameters. Then the current AMI sampling value is also put into the historical sampling sequence, and in a subsequent non-AMI sampling time period, the new fitting parameters are used to fit the data of multiple AMI historical sampling values.

[0046] Since this part is implemented in step 2, it will be described in detail later. This adjustment method is essentially to use the best fitting parameters of the previous stage to fit the next stage. Since the state change of the distribution network is not large under steady-state conditions, the fitting effect after adopting this adjustment method is better. In addition, although the invention can dynamically adjust the window length and fitting order at each sampling, when the method of the present invention samples for the first time, the two parameters of the time window length and fitting order used are pre-set. Step 2, based on the data obtained in step 1, evaluate the real-time status of the system nodes of the distribution network and output the evaluation results.

[0047] After completing the acquisition of the corresponding data in the system, the trajectory can be performed based on the acquired data and accurate evaluation results can be generated.

[0048] Preferably, in step 1, the current time is (t, t+T, t+2T, ..., t+nT), respectively; wherein T is the sampling period for sampling AMI and SCADA data in the distribution network state assessment method, t is the initial assessment time, and n is any natural number.

[0049] It is understandable that the distribution network status assessment method of the present invention will re-assess the distribution network status every T time length. In order to ensure the accuracy of each assessment, the optimal parameters used in the assessment process and the assessment results should be recorded.

[0050] Preferably, evaluating the real-time status of the system nodes of the distribution network also includes acquiring AMI and SCADA data at each current moment, and acquiring an evaluation result of the distribution network status based on the acquired data; the AMI data includes AMI sampling data and AMI pseudo-measurement data.

[0051] It is understandable that, since the evaluation time interval is set to be consistent with the sampling time interval of the SCADA system, or a set multiple of the sampling time interval of the SCADA system, the SCADA data obtained at each moment should be the latest real measurement data of the SCADA system obtained in the current period. However, since the sampling interval of AMI data is relatively large, it involves selecting the real measurement data of AMI at the sampling time of the AMI system, and generating pseudo measurement data of the AMI system at the non-sampling time of the AMI system.

[0052] Preferably, if the current moment is an AMI sampling moment, then AMI sampling data and SCADA sampling data are obtained; if the current moment is not an AMI sampling moment, then AMI pseudo-measurement data and SCADA sampling data are obtained.

[0053] It is understandable that, as described above, depending on whether the current moment is the corresponding AMI sampling moment, it is possible to choose to obtain AMI sampling data or AMI pseudo-measurement data.

[0054] Preferably, before determining the reading of SCADA sampling data and AMI sampling data, the method reads the initial network parameters of the distribution network and establishes the distribution network topology map based on the initial network parameters; the initial network parameters include the line connection relationship of the distribution network, the installation location of the SCADA equipment, and the installation location of the AMI equipment.

[0055] It can be understood that in the method of the present invention, in order to identify the content of the data actually collected by each SCADA device and each AMI device, and the relationship between each different data, so as to realize the distribution network evaluation in step 2 of the present invention, it is necessary to determine the line connection relationship of the distribution network, the network topology of the distribution network, and the actual location of each SCADA and AMI device in the distribution network before collecting data.

[0056] Therefore, the initial parameters in the present invention include information such as line connection relationship, SCADA equipment installation location, AMI equipment installation location, etc. In addition, each SCADA or AMI data acquisition device or the network router connected to the above device can realize the construction of network topology according to the above information. The construction process of this network topology is basically consistent with the content of the real-time updated routing table stored in the router in the prior art, and will not be repeated here.

[0057] Figure 2 The topological diagram of the installation location of the SCADA equipment and the AMI equipment and the distribution network architecture in the distribution network status evaluation method based on multi-section intelligent instrument data of the present invention. Figure 2 As shown, after the initial network parameters are obtained, the method of the present invention can obtain the topological structure of the distribution network. Figure 2 It can be found that in the busbars or other lines of different distribution networks, multiple lines are connected through two nodes labeled 2 and 7. Among them, 6, 5, 2, 3, and 4 are located on one busbar, 10, 9, 7, 12, and 13 are located on another busbar, and nodes 11 and 8 are located on other lines connected to the busbar.

[0058] In one embodiment of the present invention, there are three different SCADA data acquisition devices between multiple nodes of the present invention, such as between node 5 and node 2, node 2 and node 3, and node 3 and node 4, while there is no SCADA data acquisition device between node 6 and node 5, between node 9 and node 7, and between node 7 and node 8. On the other hand, the AMI data acquisition device is not installed on every node. For example, the AMI device is only installed on nodes 6, 5, 2, 8, 10, 11, 12, and 13, but not on nodes 1, 3, 4, 7, and 9.

[0059] Since the network topology diagram of the distribution network is formed before the above data is collected in the present invention, the data information collected on different nodes and between different nodes can be accurately analyzed and understood.

[0060] Preferably, the SCADA sampled data includes branch active power and branch reactive power; the AMI sampled data and the AMI pseudo-measured data include node injected active power and node injected reactive power.

[0061] It can be understood that in the method of the present invention, for SCADA equipment, the active power and reactive power on the branch where the equipment is located can be collected separately, while for AMI equipment, the active power and reactive power injected by the node on the node where the equipment is located can be collected separately.

[0062] Since the analysis method of step 2 in the present invention only needs to use active power and reactive power data to accurately evaluate and determine the status of each line of the distribution network, there is no need to collect other types of data information in the method of the present invention. However, it is easy to think of and easy to implement that in order to realize a variety of different analysis and evaluation methods, when the original data required by the evaluation method is the basic data of the power grid such as three-phase voltage and three-phase current, the SCADA equipment and AMI equipment can also easily adapt to the collection and transmission of different data. Therefore, the method of the present invention is also applicable to processing branch current amplitude measurement, node voltage amplitude measurement and node voltage complex phase measurement.

[0063] In addition, in order to evaluate the accuracy of the method, the method of the present invention can collect the topological information of the distribution network, the location of the SCADA equipment and the AMI equipment, and also collect the sampling frequency and measurement accuracy of each different SCADA equipment and AMI equipment. Through the above two data, the time and accuracy of the data collected by each intelligent meter data device can be obtained. The two parameters of data collection time and accuracy are important indicators of data, and will play an important role in the subsequent analysis and processing of data. For example, the time tag of the data can indicate whether the current data needs to be calculated at different distribution network status evaluation times, and the data accuracy can prevent the calculation redundancy of the subsequent data analysis and calculation process, and at the same time ensure that the analysis result has the highest accuracy, for example, the convergence accuracy of the status evaluation in step 2 can be set according to the accuracy of the sampled data. In one embodiment of the present invention, the standard deviation of the measurement error of the SCADA intelligent meter data device is 0.01, and the standard deviation of the measurement error of the AMI intelligent meter data device is 0.005.

[0064] In the specific analysis and calculation, since the network topology relationship can be abstracted as a node admittance matrix, the corresponding results of the state evaluation can be obtained based on the product of the data on different nodes and between nodes and the node admittance matrix.

[0065] Preferably, the updating rule steps of the sliding time window length include: step 1.1, performing one-bit increase and decrease operations on the sliding time window length and the polynomial fitting order of the previous AMI sampling moment, and permuting and combining the operation results; step 1.2, using all possible results in the permutation and combination generated in step 1.1 to fit the historical AMI sampling data, and respectively obtaining the errors between the fitting values ​​of different possible results and the sampling values ​​of the AMI sampling data at the current moment; step 1.3, selecting the possible result corresponding to the minimum error, and using the sliding time window length and the polynomial fitting order in the possible result as the fitting parameters of the AMI pseudo-measurement data between the current AMI sampling moment and the next AMI sampling moment.

[0066] It is understandable that the method of the present invention can set fitting related parameters different from or the same as those of the previous period between every two AMI sampling moments, that is, in each AMI sampling period. The two fitting parameters in the present invention are the fitting order p of the polynomial and the length s of the sliding time window.

[0067] Specifically, Figure 3 Schematic diagram of the sliding time window update in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention. Figure 3As shown, a historical AMI sampling data is placed in each basic window in the sliding time window; wherein, the historical AMI sampling data will continue to change with the addition of new AMI sampling data and the length of the sliding time window.

[0068] It is understandable that at each AMI sampling moment, a new AMI sampling data will be obtained, and at the same time, the device can store multiple historical AMI sampling data. By calculating the historical data and the new AMI sampling data, the two updated fitting parameters at the current AMI sampling moment can be obtained. The two updated fitting parameters can be used until the next AMI sampling data is collected. For one embodiment of the present invention, when the AMI sampling interval is 15 minutes, the fitting parameters will not be updated within this 15 minutes, and the fitting parameters will be updated within the next 15 minutes.

[0069] It can be understood that step 1.1 of the present invention can obtain two parameters of the previous moment, for example, moment r, and these two parameters are respectively the sliding time window length s used to generate the AMI pseudo measurement data in the time period between moment r-15min and moment r. r and polynomial fitting order p r The increase and decrease operations are performed on these two parameters. The operation method is as follows: for the sliding time window length s r , the possible values ​​of the sliding time window length for the next time period are Similarly, for the polynomial fit order p r For example, the fitting order for the next time period is

[0070] By permuting and combining the above results, we can get up to 9 different possible results, that is, the possible combinations are (s r , p r )、(s r -1, p r )、(s r +1, p r )、(s r , p r -1)、(s r -1, p r -1)、(s r +1, p r -1)、(s r , p r +1)、(s r -1, p r +1)、(s r +1, p r+1). Here, the maximum possible results are 9. However, depending on the actual situation, the values ​​of the above 9 results may not be fully realized. This is because in the actual application process, the following conditions need to be met

[0071] This condition specifically means that in order to effectively generate pseudo-measurement data for AMI, the order of fitting should be at least 1 to ensure that the right-hand side value of the polynomial regression function used for fitting will not be equal to a constant and unable to operate on the independent variable. When the fitting order is 1, the regression function can at least fit a linear function. The higher the fitting order, the higher the complexity of the curve.

[0072] In addition, in order to make the fitting order at least 1, the length of the sliding time window should be at least 1 larger than the fitting order. This is because when the fitting order is 1, one historical AMI sampling data needs to be used, and one current AMI sampling data should also be stored in the sliding window. Based on this reason, we have s r ≥p r +1 and s r ≥2.

[0073] The nine possible results are compared with the above conditions to obtain all possible results. According to each possible result, a corresponding number of historical AMI sampling data can be obtained from the sliding time window, and all possible fitting values ​​can be solved using the polynomial fitting formula using the corresponding fitting order.

[0074] In step 1.2, the formula obtained by the fitting method is:

[0075]

[0076] Among them, p is the possible fitting order used in this fitting process, k is any integer between 0 and p, t r is any time between time r and time r+15min. Its value interval can usually be obtained according to the interval of status evaluation, which is usually consistent with the sampling time of SCADA, or its corresponding multiple. r ) is the fitting result at t r AMI pseudo measurement data at the moment.

[0077] In order to realize the calculation of this formula, it is necessary to obtain the specific value of the coefficient of each term in the polynomial. In the present invention, in order to fit the optimal polynomial, the coefficient acquisition method is realized by the least square method.

[0078] However, in the present invention, it should be noted that the original data used by the least squares method are multiple AMI sampling times t i Multiple different AMI sampling data on the same time frame are real AMI data. i and t i+1 ,t i-1 Therefore, the present invention actually fits the best polynomial fitting curve over a very long period of time, and only applies the current fitting to one of ti and ti. +1 The time interval is between time r and time r+15min.

[0079] Specifically, the number of AMI sampling data used by the least squares method is determined by the length of the sliding time window, that is, the formula of the least squares method in the present invention is: Where s is the length of the sliding time window, which is also the number of historical AMI sampling data used by the least squares method in the present invention. Since the least squares method is also a common content in the prior art, the least squares method for the parameter a in the formula will not be described here. p 、a p-1 The calculation process to a0.

[0080] The least squares method is used to obtain f(t r ) formula, the current moment can be substituted into the formula to obtain the fitting value. The fitting values ​​calculated by the 9 different possible results are all different. Therefore, in the present invention, multiple fitting values ​​are compared with the AMI sampling data at the current moment to obtain error values. If all 9 possible results are valid, 9 error values ​​will be generated accordingly. The smallest of the 9 error values, that is, min{e1, ..., e9}, is obtained to obtain the optimal fitting parameters.

[0081] That is to say, in the next time period, the optimal fitting parameters can be used to generate AMI pseudo-measurement data.

[0082] The method adopted by the present invention is suitable for the case where the distribution network is working in a stable state and the power grid current does not change much, and an accurate assessment of the state of the distribution network is achieved by gradually adjusting the fitting parameters with a small amplitude. Since the method of the present invention generates high-precision AMI pseudo-measurement data to make the distribution network in the non-AMI sampling period observable, the execution cycle of the state assessment is shortened, and its accuracy is higher than the traditional method of directly using the AMI sampling value closest in time as the constant pseudo-measurement in the subsequent non-AMI sampling period, thereby providing more real-time and accurate data for the operation control of the distribution network.

[0083] It should be noted that, in addition to the least square method commonly used in the prior art, the present invention can also use an improved least square method, that is, a weighted least square method, to perform state evaluation processing on all measurement data. When the weighted least square method is used, the node voltage complex phasor can be selected as the state variable to generate the state vector x = [x1, x2, ..., x n ] T , using a variety of measurement data to form a measurement vector z = [z1, z2, ..., z m ] T Under the conditions of given grid connection, branch parameters and measurement system, the relationship between the state vector and the measurement vector in the system can be obtained as z = h(x) + v. Among them, h(x) is the nonlinear measurement function of the state vector, and v is a function with a mean of 0 and a normal distribution. Gaussian white noise, where is the white noise variance of the i-th quantity measurement.

[0084] The objective function of weighted least squares state assessment is minJ(x)=min[zh(x)] T R -1 [zh(x)], where R -1 is the measurement weight matrix of the system. By finding the minimum value of the objective function, the nonlinear measurement equation in the present invention can be obtained. Since the weighted least square method is a common method in the prior art, it will not be described here.

[0085] Preferably, initialization is performed before obtaining the AMI pseudo-measurement data, and the initial values ​​of the sliding time window length and the polynomial fitting order are set during the initialization process; when the weighted least squares method is used, the initial values ​​of the state variables and the convergence accuracy of the state evaluation are set during the initialization process.

[0086] It should be noted that in the process of obtaining the best fitting function by the least square method and calculating the pseudo measurement data by the fitting function, the system implementing the method of the present invention needs to perform a system initialization operation after receiving the AMI measurement data. During the initialization operation, the initial values ​​of various parameters need to be set first.

[0087] Preferably, the range of the initial value s1 of the sliding time window length is [2, 10], and the range of the initial value p1 of the polynomial fitting order is [1, 5]; and the parameters satisfy the condition s1 ≥ p1 + 1. This is due to the parameter characteristics of the two related parameters of the sliding time window length and the polynomial fitting order. This part of the content has been explained above and will not be repeated here.

[0088] Specifically, if the weighted least square method is used for state evaluation in the present invention, the initial value of the state variable needs to be set. In the process of state evaluation, in order to ensure that the convergence accuracy meets the requirements, the data acquisition accuracy corresponding to the device is usually obtained according to the parameters of each device, and the convergence accuracy of the state evaluation is set according to the data acquisition accuracy. The convergence accuracy of the state evaluation in the present invention can be set in [10 -6 , 10 -4 ]between.

[0089] At the same time, no matter which fitting function acquisition method is adopted, in order to ensure the smooth execution and termination of the method of the present invention, the size of the state evaluation period T and the termination time of the state evaluation are usually set. As mentioned above, the size of the state evaluation period T is usually much smaller than the time length of the AMI sampling period of 15 minutes, and should be an integer multiple of the SCADA data sampling period. In one embodiment of the present invention, the time T is 1 minute.

[0090] In the present invention, the accuracy of pseudo-measurement data can be tested by using the mean absolute error method. The calculation formula of the mean absolute error in the present invention is:

[0091]

[0092]

[0093] Wherein, N is the number of system nodes in the distribution network, and M is the total number of state evaluations in the method of the present invention, and its value is related to the end time and period T of the state evaluation set in the initialization process. In addition, is the active power, Can be pseudo measurement data The measurement value corresponding to the i-th state evaluation node at the j-th evaluation; is the reactive power, The pseudo measurement data The reactive power measurement value corresponding to the i-th state evaluation node at the j-th evaluation.

[0094] In order to verify the effect of the present invention, three different methods are selected in the present invention to compare the results obtained by different algorithms.

[0095] The first method is to use the commonly used method in the prior art to achieve the mixed measurement of SCADA and AMI. In this method, the measurement is only performed within the AMI sampling time interval of 15 minutes, and the status evaluation is performed based on the measurement results.

[0096] The second method is to use the pseudo measurement data in the present invention to realize the generation of pseudo measurement data at non-AMI sampling time. On this basis, the SCADA data and AMI data are integrated to realize the state evaluation. In this embodiment, the least square method is selected as the method for obtaining the fitting function, and the initial value of the sliding time window length is set to 6, and the initial value of the polynomial fitting order is 2.

[0097] The third method adopts the method in the prior art, and directly adopts the AMI sampling value closest in time as the constant pseudo measurement value in the non-AMI sampling time period, and then performs state evaluation.

[0098] Figure 4 The figure is a schematic diagram showing the changes of the two parameters of the basic window number and the fitting order over time in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention. Figure 4 As shown, in the second method, although the initial parameters are 6 and 2 respectively, as time changes, with the calculation of the fitting function error and the iterative adjustment of the above two parameters, the parameter values ​​change in multiple state evaluation cycles. For example, the fitting order takes a value of +1 in the third evaluation and +1 again in the fourth evaluation, while the window length is reduced to 5 in the fourth fitting.

[0099] Table 1 is a comparative data table of pseudo measurement accuracy of the second method and the third method of the present invention. As shown in Table 1, the accuracy of the AMI pseudo measurement data of the second method is higher than the pseudo measurement data generation method of the third method.

[0100] Way Active power error value (pu) Reactive power error value (pu) Second method 0.00205 0.00121 The third way 0.00293 0.00185

[0101] Table 1 Comparison of pseudo-measurement accuracy between the second method and the third method

[0102] Figure 5 It is a schematic diagram for comparing the AMI pseudo-measurement data obtained in the second mode of the distribution network state assessment method based on multi-section intelligent instrument data of the present invention with the real injected power of the distribution network. Figure 6 This is a schematic diagram of the comparison between the AMI pseudo-measurement data obtained in the third method of the distribution network state assessment method based on multi-section intelligent instrument data of the present invention and the real injected power of the distribution network. It can be clearly seen from the figure that the value of the pseudo-measurement data obtained in the second method is closer to the real measurement data than the third method.

[0103] In the present invention, in order to facilitate quantitative analysis of the state assessment accuracy, the maximum absolute error of voltage amplitude, the maximum absolute error of phase angle, the average absolute error of voltage amplitude and the average absolute error of phase angle can be used as measurement indicators. The error expressions obtained according to the above indicators are:

[0104]

[0105]

[0106]

[0107]

[0108] in, V is the estimated value of the voltage amplitude of node i in the jth state evaluation; i j Evaluate the true value of the voltage amplitude at node i for the jth state; is the estimated value of the voltage phase angle of node i in the jth state evaluation; δ i j Evaluate the true value of the voltage phase angle at node i for the jth state.

[0109] Table 2 shows the various error indicators of the state evaluation of node 7 under the above three methods of the present invention. As shown in Table 2, for the current node, since the first method does not generate pseudo measurement data, its average absolute error is the smallest. However, although the second method generates a large amount of pseudo measurement data, its average absolute error is not much larger than that of the first method, and its average absolute error is also smaller than that of the third method.

[0110]

[0111] Table 2 Comparison of various error indicators of node 7 in state evaluation under three methods

[0112] Figure 7 It is a schematic diagram for comparing the average errors of node three-phase voltage amplitudes in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention. Figure 8 It is a schematic diagram for comparing the average errors of the phase angles of the three-phase voltages at nodes in a distribution network state assessment method based on multi-section intelligent instrument data of the present invention. Figure 7 and Figure 8 The comparison of the average errors of the voltage amplitude and the voltage phase angle in the three methods of the present invention are respectively shown. The error of the second method is relatively close to that of the first method, and is much smaller than the error of the third method. It can be seen that the method in the present invention well realizes the prediction and generation of relevant data in the non-AMI sampling time period, and the pseudo-measurement accuracy is higher, so that the final state evaluation accuracy is higher than the method of selecting the AMI sampling value closest to the sampling time as the pseudo-measurement. At the same time, the present invention can perform multiple effective evaluations within an AMI sampling cycle, so the present invention also effectively shortens the execution cycle of the distribution network state evaluation.

[0113] The beneficial effect of the present invention is that, compared with the prior art, the distribution network status method based on multi-section intelligent instrument data in the present invention collects AMI sampling data and SCADA sampling data to evaluate the system node status of the distribution network in real time, and extracts the most suitable AMI pseudo-measurement data by dynamically adjusting the sliding window length and the polynomial order, thereby realizing high-precision distribution network status evaluation. The method of the present invention has a simple calculation process, the pseudo-measurement results can be flexibly and smoothly fine-tuned based on historical data, and the evaluation results are accurate.

[0114] The applicant of the present invention has made a detailed explanation and description of the implementation examples of the present invention in conjunction with the drawings in the specification. However, those skilled in the art should understand that the above implementation examples are only preferred implementation schemes of the present invention, and the detailed description is only to help readers better understand the spirit of the present invention, but not to limit the scope of protection of the present invention. On the contrary, any improvements or modifications based on the inventive spirit of the present invention should fall within the scope of protection of the present invention.

Claims

1. A distribution network status assessment method based on smart meter data, characterized in that: The method comprises the following steps: Step 1, obtaining the AMI sampling time, and determining the reading of SCADA sampling data and AMI sampling data based on the comparison between the current time and the AMI sampling time; When the AMI sampling data cannot be read, the AMI pseudo-measurement data is fitted using a polynomial fitting method based on the preset sliding time window length and polynomial fitting order; Each window in the sliding time window contains a historical AMI sampling data or a current AMI sampling data; The historical AMI sampling data will continue to change with the addition of new AMI sampling data and the length of the sliding time window; The calculation process of the sliding time window length and the polynomial fitting order includes: Step 1.1, perform one-bit increase or decrease operations on the sliding time window length and polynomial fitting order of the previous AMI sampling moment, and perform permutation and combination of the operation results; Among them, for the previous sliding time window length s r and the fitting order p of the previous time period r Using one-bit increase and decrease operations for calculation, the possible values ​​of the sliding time window length for the next time period are: Possible values ​​for the fit order for the next time period are: Generates r+1 With p r+1 All possible results of the permutations and combinations of All possible results are (s r ,p r )、(s r -1,p r )、(s r +1,p r )、(s r ,p r -1)、(s r -1,p r -1)、(s r +1,p r -1)、(s r ,p r +1)、(s r -1,p r +1)、(s r +1,p r +1) one or more; And all possible results satisfy the condition Step 1.2, fitting the historical AMI sampling data using all possible results in the permutation and combination generated in step 1.1, and obtaining the errors between the fitting values ​​of different possible results and the sampling values ​​of the AMI sampling data at the current moment; Step 1.3, select the possible result corresponding to the minimum error, and use the sliding time window length and polynomial fitting order in the possible result as the fitting parameters of the AMI pseudo-measurement data between the current AMI sampling time and the next AMI sampling time; Step 2: Based on the data obtained in step 1, the real-time status of the system nodes of the distribution network is evaluated and the evaluation results are output.

2. A distribution network status assessment method based on smart meter data according to claim 1, characterized in that: In step 1, the current time is (t, t+T, t+2T, ..., t+nT) respectively; wherein, T is the sampling period for sampling AMI and SCADA data in the distribution network state evaluation method, t is the initial evaluation time, and n is an arbitrary natural number.

3. A distribution network status assessment method based on smart meter data according to claim 2, characterized in that: T is an integer multiple of the sampling interval of the SCADA sampling data, and T is much smaller than the sampling interval of the AMI sampling data.

4. A distribution network status assessment method based on smart meter data according to claim 3, characterized in that: The evaluating of the real-time status of the system nodes of the distribution network also includes acquiring AMI and SCADA data at each current moment, and acquiring an evaluation result of the distribution network status based on the acquired data; The AMI data includes AMI sampling data and AMI pseudo-measurement data.

5. A distribution network status assessment method based on smart meter data according to claim 4, characterized in that: If the current time is the AMI sampling time, then obtain the AMI sampling data and SCADA sampling data; If the current time is not the AMI sampling time, the AMI pseudo measurement data and SCADA sampling data are obtained.

6. A distribution network status assessment method based on smart meter data according to claim 5, characterized in that: The method reads initial network parameters of the distribution network before determining the reading of SCADA sampled data and AMI sampled data, and establishes the distribution network topology map based on the initial network parameters; The initial network parameters include the line connection relationship of the distribution network, the installation location of the SCADA equipment, and the installation location of the AMI equipment.

7. A distribution network status assessment method based on smart meter data according to claim 6, characterized in that: The SCADA sampling data includes branch active power and branch reactive power; The AMI sampling data and the AMI pseudo-measurement data include node-injected active power and node-injected reactive power.

8. A distribution network status assessment method based on smart meter data according to claim 7, characterized in that: The fitting parameters of the AMI pseudo-measurement data between the current AMI sampling moment and the next AMI sampling moment obtained in step 1.3 are used to generate corresponding AMI pseudo-measurement data at each current moment to achieve state evaluation of the distribution network.

9. A distribution network status assessment method based on smart meter data according to claim 8, characterized in that: The fitting function is solved using the least squares method or the weighted least squares method.

10. A distribution network status assessment method based on smart meter data according to claim 9, characterized in that: When the weighted least squares method is used, the node voltage complex phasor is taken as the state variable x= [x1,x2,…,x n ] T , using the measured data as the measurement vector z = [z1, z2, …, z m ] T , where z = h(x) + v, the calculation meets the objective function minJ(x) = min[zh(x)] T R -1 [zh(x)] The value of x under the condition; Where h(x) is the nonlinear measurement function of the state vector, v is Gaussian white noise, R -1 is the measurement weight matrix.

11. A distribution network status assessment method based on smart meter data according to claim 10, characterized in that: Initialization is performed before acquiring the AMI pseudo-measurement data, and initial values ​​of the sliding time window length and the polynomial fitting order are set during the initialization process.

12. A distribution network status assessment method based on smart meter data according to claim 11, characterized in that: When the weighted least square method is used, the initial value of the state variable and the convergence accuracy of the state evaluation are set during the initialization process.

13. A distribution network status assessment method based on smart meter data according to claim 12, characterized in that: The initial value s1 of the sliding time window length is in the range of [2, 10], and the initial value p1 of the polynomial fitting order is in the range of [1, 5]; And the parameters satisfy the condition s1≥p1+1.

14. A distribution network status assessment method based on smart meter data according to claim 13, characterized in that: Mean absolute error and Verify the accuracy of pseudo-measurement data; Wherein, N is the number of system nodes in the distribution network, M is the number of state evaluations, which depends on the end time of the state evaluation and the period T set in the initialization process. P is the active power, P i j The pseudo measurement data The active power measurement value corresponding to the i-th state evaluation node at the j-th evaluation; Q is reactive power, The pseudo measurement data The reactive power measurement value corresponding to the i-th state evaluation node at the j-th evaluation.

15. A distribution network status assessment method based on smart meter data according to claim 14, characterized in that: When the sampling period of the SCADA data is 1 s and the sampling period of the AMI data is 15 min, the sampling period T for sampling the AMI and SCADA data is set to 1 min.

16. A distribution network status assessment method based on smart meter data according to claim 15, characterized in that: Using mean absolute error to test the accuracy of the pseudo-measurement data; The accuracy of the evaluation result is tested using the maximum absolute error of the voltage amplitude, the maximum absolute error of the voltage phase angle, the average absolute error of the voltage amplitude and the average absolute error of the voltage phase angle of the pseudo-measurement data.

Images