Static power distribution network system observability evaluation method based on random response surface method

Abstract

The invention provides a static power distribution network system observability evaluation method based on a random response surface method, and the method comprises the steps: on the basis of carrying out the observability analysis of a static power distribution network system through employing the random response surface method, judging whether the power distribution network system is observable or not: when and only when an observability coefficient matrix has n linear independent column vectors, carrying out the observability analysis of the static power distribution network system; the system is observable, and the ith row and the (n + i) th row cannot be zero vectors at the same time. According to the method, an index which can still perform observability analysis on the static power distribution network system under the scenes that the measurement data contains noise, the measurement data is missing and the measurement data is delayed is further constructed, so that normal and stable operation of the power distribution network system is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution network systems, and in particular to an observability evaluation method for a static power distribution network system based on a random response surface method. Background Art

[0002] The observability of the power system is a prerequisite for state estimation calculation, so it has always been the focus of research. The observability analysis and observability quantification of the distribution network system play a vital role in the study of modern power systems. The observability analysis methods of static distribution network systems include static state estimation, least squares method and topological method, but the applicability of observability analysis methods based on static state estimation is based on the premise that the measurement data is completely accurate, and cannot be applied to noisy data, lost data or delayed data in practical applications; the least squares method is sensitive to abnormal data and noise, which may lead to erroneous state estimation; the topological method is used for large-scale and complex distribution network systems, and the computational complexity is high. The observability quantification indicators of static distribution network systems include observability measurement matrix, information theory index and redundancy measurement, but the observability measurement matrix lacks robustness to data noise; the information theory index has poor performance in high noise environment, has high requirements on data quality, and does not allow any data loss or delay; the redundancy measurement requires a large number of measurement points to increase system cost, and too much redundancy may lead to complexity of data processing and waste of resources.

[0003] Therefore, the existing observability analysis methods and observability quantification indicators applied to static distribution network systems cannot resist the interference of noise, are not suitable for scenarios where data is lost or delayed, and data processing is more complicated. The measurement data in actual projects contains noise, and the measurement data may be lost or delayed due to sensor failure and other reasons. In view of the fact that the existing observability analysis methods and observability quantification indicators applied to static distribution network systems are completely unable to accurately obtain the observability of the system in the above scenarios. Therefore, it is necessary to consider the interference of noise and the loss and delay of measurement data in the observability analysis methods and observability quantification indicators of static distribution network systems. Summary of the invention

[0004] In view of the defects and shortcomings of the prior art, the purpose of the present invention is to provide a method for evaluating the observability of a static distribution network system based on the random response surface method, and to provide a method for determining whether the system is observable. The method takes into account the randomness of the system, and the method has low computational complexity; and further considers the scenarios where the measurement data is lost and delayed, and proposes corresponding quantitative observability indicators; finally, the observability conditions and comprehensive observability indicators are combined, and a set of solutions for determining the observability of the system is proposed.

[0005] The present invention specifically adopts the following technical solutions:

[0006] The present invention first provides a method for judging the observability of a static distribution network system based on a random response surface method. On the basis of using the random response surface method to perform observability analysis on the static distribution network system, the basis for judging whether the distribution network system is observable is: if and only if the observability coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

[0007] Furthermore, the observability coefficient matrix is ​​obtained by:

[0008] The random response surface method uses a truncated expanded surrogate model:

[0009] The random output y is:

[0010]

[0011] Where φ0, φ1(ξ i ),φ2(ξ i 2 ) represent the zero-order, first-order, and second-order random response surface method basis, a0, a i , a i,i represents the corresponding random response surface method coefficient;

[0012] Zero-order random response surface method basis and its corresponding coefficients φ0 and represents the average of the estimated values ​​of the lth measurement, as well as the first-order and second-order random response surface method bases and their corresponding coefficients and represents the uncertainty of the estimated value of the lth measurement relative to the i-th state estimate, where the random response surface method coefficient and represents the contribution of the i-th state to the uncertainty of the estimated value of the l-th measurement;

[0013] set up represents the observability coefficient matrix, which is a submatrix of the coefficient matrix with n linearly independent columns; as follows:

[0014]

[0015] On this basis, the observability evaluation method of the static distribution network system based on the random response surface method of the present invention is constructed:

[0016] On the basis of judging the observability of the static distribution network system by the above method, the strength of the observability of the distribution network system is quantified by one or more of the following indicators, namely: the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system.

[0017] Among them, the above five indicators can objectively characterize the strength of the observability of the distribution network system on the basis of observability. Each of them is aimed at a different observability angle. In the specific application process, technical personnel in this field can select one or more of them to evaluate the strength of observability according to the actual situation. Of course, if one or more of them is reduced, the overall evaluation may be imperfect, but in certain scenarios with clear conditions (such as obviously weak noise, no data loss or clear low-latency scenarios), partial evaluation is still effective. However, the comprehensive evaluation of the five indicators is still the most complete and preferred solution of the present invention. For this reason, the present invention provides the following preferred design.

[0018] Furthermore, the method to quantify the strength of the observability of the distribution network system is as follows:

[0019] Constructing a comprehensive index R to quantify the impact of observability total

[0020]

[0021] Among them, R1~R5 are the observability quantitative indicators corresponding to the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system;

[0022] R total The value of reflects the strength of the system observability. The closer it is to 1, the stronger the system observability is; the closer it is to 0, the weaker the system observability is.

[0023] Let R o represents the rank of the observability coefficient matrix; R o =n, the system is observable, which is the premise for analyzing the observability of the system;

[0024] The observability condition R o Combined with observability, a comprehensive index R that quantifies the degree of observability impact total , construct an index R that can quantify the degree of observability of the system all ;

[0025] R all =R o +R total ,Rall ∈(n,n+1)

[0026] When R all ∈(n,n+0.35), the system is weakly observable; when R all ∈(n+0.35,n+0.75), the observability of the system is moderate; when R all ∈(n+0.75,n+1), the system is strongly observable;

[0027] If the observability condition R o <n, the system is unobservable.

[0028] Furthermore, the condition number of the matrix is ​​the ratio of the largest singular value to the smallest singular value, that is:

[0029]

[0030] The reciprocal of the condition number is used as the observability quantitative indicator R1:

[0031]

[0032] The closer R1 is to 0, the weakly observable system is; the closer R1 is to 1, the observable system is.

[0033] Furthermore, the observability of the distribution network system is quantified by the contribution of the state to the estimation:

[0034] According to the proxy model, the variance of the estimated value of the lth measurement is:

[0035]

[0036] The ratio of the contribution of the i-th state to the variance of the estimated value of the l-th measurement is defined as the contribution ratio,

[0037]

[0038] in,

[0039] The observability quantitative index R2 is constructed as:

[0040]

[0041] If the system is strongly observable, R2 is close to 1; if the system is weakly observable, R2 is close to 0.

[0042] Furthermore, the observability of the distribution network system is quantified by observing the impact of noise on the observability of the system:

[0043] The ratio of the noise variance of the lth measurement to the variance of the estimated value is defined as the interference rate:

[0044]

[0045] The observability quantitative index R3 is constructed as:

[0046]

[0047] in, Indicates the average interference rate of the measurement;

[0048] If the system is weakly observable, R3 is close to 0.

[0049] Furthermore, the observability of the distribution network system is quantified by observing the impact of data loss on the observability of the system:

[0050] is the estimated variance of the lth measurement when no data is missing, is the variance of the estimated value after the data loss occurs; represents the variance of the estimated value of the lth measurement due to missing observation data;

[0051] The ratio of the difference in variance before and after the observed data loss of the lth measurement to the variance of the estimated value when no data loss occurs is defined as the data loss impact factor:

[0052]

[0053] The observability quantitative index R4 is constructed as:

[0054]

[0055] in, Indicates the average data loss impact factor of the measurement;

[0056] If the system is weakly observable, R4 is close to 0.

[0057] Furthermore, the observability of the distribution network system is quantified by observing the impact of data delay on the observability of the system:

[0058] is the estimated value variance of the lth measurement without data delay, is the variance of the estimated value after the data delay occurs; represents the variance of the estimated value of the lth measurement due to the delay of the observed data;

[0059] The ratio of the difference between the variances of the observed data before and after the data delay and the variance of the estimated value without data delay is defined as the data delay impact factor.

[0060]

[0061] The observability quantitative index R5 is constructed as follows:

[0062]

[0063] in, Indicates the measured average data delay impact factor;

[0064] If the system is weakly observable, R5 is close to 0.

[0065] The present invention also requests protection for a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores a computer program capable of running on the processor. When the processor loads and executes the computer program, a static distribution network system observability assessment method based on the random response surface method as described above is adopted, which at least includes the above-mentioned observability judgment module and preferably includes a quantitative assessment module.

[0066] And, a static distribution network system observability evaluation device based on random response surface method, including an observability judgment module, which judges whether the distribution network system is observable based on the observability analysis of the static distribution network system using the random response surface method; the basis is: if and only if the observability coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

[0067] Furthermore, it also includes a quantitative evaluation module: based on the observability judgment module, the observability of the static distribution network system is judged to be observable, and one or more of the following indicators are used to quantify the strength of the observability of the distribution network system, namely: the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system.

[0068] Furthermore, the quantitative evaluation module uses a comprehensive index R to quantify the degree of observability impact. total

[0069]

[0070] Among them, R1~R5 are the observability quantitative indicators corresponding to the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system;

[0071] R totalThe value of reflects the strength of the system observability. The closer it is to 1, the stronger the system observability is; the closer it is to 0, the weaker the system observability is.

[0072] Let R o represents the rank of the observability coefficient matrix; R o =n, the system is observable, which is the premise for analyzing the observability of the system;

[0073] The observability condition R o Combined with observability, a comprehensive index R that quantifies the degree of observability impact total , construct an index R that can quantify the degree of observability of the system all ;

[0074] R all =R o +R total ,R all ∈(n,n+1)

[0075] When R all ∈(n,n+0.35), the system is weakly observable; when R all ∈(n+0.35,n+0.75), the observability of the system is moderate; when R all ∈(n+0.75,n+1), the system is strongly observable;

[0076] If the observability condition R o <n, the system is unobservable.

[0077] Furthermore, the condition number of the matrix is ​​the ratio of the largest singular value to the smallest singular value, that is:

[0078]

[0079] The reciprocal of the condition number is used as the observability quantitative indicator R1:

[0080]

[0081] The closer R1 is to 0, the weakly observable system is; the closer R1 is to 1, the observable system is.

[0082] Furthermore, the observability of the distribution network system is quantified by the contribution of the state to the estimation:

[0083] According to the proxy model, the variance of the estimated value of the lth measurement is:

[0084]

[0085] The ratio of the contribution of the i-th state to the variance of the estimated value of the l-th measurement is defined as the contribution ratio,

[0086]

[0087] in,

[0088] The observability quantitative index R2 is constructed as:

[0089]

[0090] If the system is strongly observable, R2 is close to 1; if the system is weakly observable, R2 is close to 0.

[0091] Furthermore, the observability of the distribution network system is quantified by observing the impact of noise on the observability of the system:

[0092] The ratio of the noise variance of the lth measurement to the variance of the estimated value is defined as the interference rate:

[0093]

[0094] The observability quantitative index R3 is constructed as:

[0095]

[0096] in, Indicates the average interference rate of the measurement;

[0097] If the system is weakly observable, R3 is close to 0.

[0098] Furthermore, the observability of the distribution network system is quantified by observing the impact of data loss on the observability of the system:

[0099] is the estimated variance of the lth measurement when no data is missing, is the variance of the estimated value after the data loss occurs; represents the variance of the estimated value of the lth measurement due to missing observation data;

[0100] The ratio of the difference in variance before and after the observed data loss of the lth measurement to the variance of the estimated value when no data loss occurs is defined as the data loss impact factor:

[0101]

[0102] The observability quantitative index R4 is constructed as:

[0103]

[0104] in, Indicates the average data loss impact factor of the measurement;

[0105] If the system is weakly observable, R4 is close to 0.

[0106] Furthermore, the observability of the distribution network system is quantified by observing the impact of data delay on the observability of the system:

[0107] is the estimated value variance of the lth measurement without data delay, is the variance of the estimated value after the data delay occurs; represents the variance of the estimated value of the lth measurement due to the delay of the observed data;

[0108] The ratio of the difference between the variances of the observed data before and after the data delay and the variance of the estimated value without data delay is defined as the data delay impact factor.

[0109]

[0110] The observability quantitative index R5 is constructed as follows:

[0111]

[0112] in, Indicates the measured average data delay impact factor;

[0113] If the system is weakly observable, R5 is close to 0.

[0114] Compared with the prior art, the present invention and its preferred solution firstly propose an observability analysis method based on the random response surface method, and provide a method for judging whether the distribution network system is observable: the system is observable if and only if the observable coefficient matrix has n linearly independent column vectors. And further propose an observability quantitative index based on the random response surface method to achieve quantitative evaluation of the observability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0116] Figure 1 It is a schematic diagram of the observability evaluation process of an embodiment of the present invention. DETAILED DESCRIPTION

[0117] Hereinafter, the specific embodiments of the present application will be described in detail with reference to the accompanying drawings. According to these detailed descriptions, those skilled in the art can clearly understand the present application and can implement the present application. Without violating the principles of the present application, the features in the various embodiments can be combined to obtain new implementations, or certain features in certain embodiments can be replaced to obtain other preferred implementations.

[0118] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail with the accompanying drawings as follows:

[0119] 1. Principles of random response surface methodology

[0120] (1) Mapping transformation of random input

[0121] According to the following standard probability distribution, the i-th state random variable x i Mapped to a new random variable ξ i :

[0122] x i =F i -1 (T i (ξ i )),i=1,2,...,n, (1)

[0123] where F i -1 (.) represents x i The inverse cumulative distribution function, T i Represents ξ i The cumulative distribution function of .

[0124] (2) Function approximation of system output

[0125] In the random response surface method, the random output y can be approximated as:

[0126]

[0127] Where y is the system output, ξ=[ξ1,ξ2,...,ξ n ] is a vector of random variables that follow a standard probability distribution, and its corresponding random response surface method basis is φ i (ξ), a i is the i-th random response surface method coefficient,

[0128] n p =(n+p)! / (n!p!)-1, and p is the maximum order of the random response surface method basis function.

[0129] From the random response surface method coefficients, the mean and variance of the output can be directly obtained as follows:

[0130]

[0131] (3) Proxy Model

[0132] In order to model the response surface with low computational complexity while maintaining high accuracy, this embodiment adopts truncated expansion, which is:

[0133]

[0134] Where φ0, φ1(ξ i ),φ2(ξ i 2 ) represent the zero-order, first-order, and second-order random response surface method basis, a0,

[0135] a i , a i,i represents the corresponding random response surface method coefficients.

[0136] 2. Observability analysis method based on random response surface methodology

[0137] To implement the surrogate model, some configuration points are selected by combining a finite sample of ξ. Considering N combinations of configuration points, the polynomial chaotic basis can be directly obtained, and the system output can be calculated through the system model. Formally, the surrogate model has

[0138] Y k =HA k (5)

[0139]

[0140]

[0141] in, is a measurement matrix consisting of the measurement values ​​of N samples; is the basis matrix composed of the random response surface method basis; ξ s,i is the i-th element of the s-th sample; is the coefficient matrix; represents the random response surface method coefficients with respect to the ith state and the lth measurement.

[0142] Assume that the random variable ξ i follows a Gaussian distribution. In this case, the corresponding Hermite polynomials are φ0=1, φ1(ξ s,i )=ξ s,i and Therefore, the elements of the collocation points are determined by the union of the zeros and roots of a high-order one-dimensional Hermitian polynomial. Then, through the basis matrix and the measurement matrix, the coefficient matrix A can be obtained k ,

[0143] A k =H -1 Y k (8)

[0144] Based on the random response surface method, the present invention proposes an observability condition to test whether the system is observable. Due to the orthogonal property, the average value of the first-order and second-order random response surface method basis is zero, so the zero-order random response surface method basis and its corresponding coefficients φ0 and represents the average of the estimated values ​​of the lth measurement, as well as the first-order and second-order random response surface method bases and their corresponding coefficients and represents the uncertainty of the estimated value of the lth measurement relative to the i-th state estimate, where the random response surface method coefficient and It represents the contribution of the i-th state to the uncertainty of the estimated value of the l-th measurement.

[0145] If the contribution of the i-th state to the uncertainty of the estimate at the l-th measurement is zero, then the latter remains unchanged with respect to changes in this state, that is, the i-th state cannot be inferred from the l-th measurement. In order to uniquely infer n states from a given measurement, n valid measurements are required for which the contributions of the states to the uncertainty of the estimates of these measurements are linearly independent. Let represents the observability coefficient matrix, which is a submatrix of the coefficient matrix with n linearly independent columns. It is as follows:

[0146]

[0147] Method to determine whether the distribution network system is observable: if and only if the observable coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

[0148] 3. Observability Quantitative Indicators Based on Stochastic Response Surface Methodology

[0149] Considering the static distribution network system, observability analysis is a key step to ensure that the system state can be accurately estimated with limited measurement data. However, in practical applications, the noise, loss and delay of measurement data have a significant impact on the observability of the system. These problems not only affect the accuracy and reliability of system state estimation, but also have a profound impact on the operation, monitoring and management of the distribution network.

[0150] Noise is an inevitable error in measurement data, which mainly comes from the inherent error of the measurement equipment, environmental interference, and loss and interference in the signal transmission process. Noise will reduce the accuracy of the measurement data, resulting in an increase in the error of the system state estimation, making the state estimation result inaccurate, thus affecting the accuracy of observability analysis; noise introduces additional uncertainty, and excessive uncertainty will reduce the observability of the system, making it difficult to accurately judge the actual operating state of the system; noise will affect the effect of fusion and filtering, resulting in large errors in the fused data; noise will interfere with the calculation of observability quantitative indicators and affect the rank calculation of the observability metric matrix, resulting in inaccurate observability analysis results.

[0151] Data loss refers to the loss of some measured data due to various reasons during the data transmission process. Data loss will lead to incomplete data sources for system state estimation, making the estimation of system state inaccurate and affecting the results of observability analysis; data loss will reduce the reliability of the system, especially data loss at key nodes, which will have a significant impact on the operation of the entire distribution network. Data loss will lead to inaccurate calculation of observability quantitative indicators and affect the overall evaluation of the system.

[0152] Delay refers to the time delay from the acquisition to the processing and analysis of measured data. Delay will affect the real-time monitoring and response speed of the system, making it difficult to estimate and control the system state in real time, reducing the observability of the system; delay will interfere with the multi-sensor data fusion and filtering process, resulting in the fused data being unable to accurately reflect the actual state of the system, affecting the results of observability analysis; delay will affect the real-time calculation of observability quantitative indicators, making these indicators unable to accurately reflect the current system state, thus affecting the evaluation and decision-making of the system.

[0153] In summary, the embodiments of the present invention further construct indicators that can still perform observability analysis on the static distribution network system in scenarios where the measurement data contains noise, measurement data is missing, and measurement data is delayed, so as to ensure the normal and stable operation of the distribution network system.

[0154] However, the observability evaluation of the existing distribution system is generally limited to whether the system is observable, and the degree of observability is not quantified. Therefore, based on the observability conditions, the present invention proposes an index to quantify the degree of observability, that is, to quantify the strength of the observability of the distribution network system. The following five quantitative indicators are proposed:

[0155] (1) Condition number

[0156] In order to ensure the numerical stability of the estimated state, the singular values ​​of the observability coefficient matrix cannot differ too much, that is, the condition number cannot be too large, otherwise the observability coefficient matrix may be ill-conditioned. If the observability coefficient matrix is ​​ill-conditioned, it is impossible to perform observability analysis on the system through the observability coefficient matrix, that is, the system is not observable. The condition number of a matrix is ​​the ratio of its maximum singular value to its minimum singular value, that is,

[0157]

[0158] The first observability quantitative index R1 is the inverse of the condition number, that is,

[0159]

[0160] Definition 1: If the condition number is very large or becomes infinite, that is, the smaller R1 is, the closer it is to 0, then the system is weakly observable. If the condition number is close to 1, that is, the larger R1 is, the closer it is to 1, then the system is observable.

[0161] (2) Contribution of state to estimation

[0162] According to the surrogate model, the variance of the estimate of the lth measurement is given by

[0163]

[0164] The ratio of the contribution of the i-th state to the variance of the estimated value of the l-th measurement is defined as the contribution ratio,

[0165]

[0166] It follows The contribution ratio indicates the influence of the state on the estimated value of the measurement. A larger contribution ratio means that the state has a greater influence on the estimated value of the measurement, and vice versa.

[0167] Definition 2: If all contribution rates If the probability of the ith state of is less than a small positive value (say 0.1%), the state is weakly observable. Otherwise, it is strongly observable. If all states are strongly observable, then the system is strongly observable. Otherwise, the system is weakly observable.

[0168] The second observability quantitative indicator R2 is,

[0169]

[0170] If the system is strongly observable, the contribution rates of all states are relatively large and close to 1, so the index R2 after taking the average value is also large and close to 1; if the system is weakly observable, not all state contribution rates are close to 1, so the value of the index R2 after taking the average value is not large and close to 0. Therefore, the size of the index R2 can reflect the strength of the observability of the system.

[0171] (3) The impact of observation noise on system observability

[0172] if The contribution of the i-th state to the variance of the estimate of the l-th measurement is close to or less than the measurement noise variance Its impact is difficult to distinguish from the impact of noise variance.

[0173] The ratio of the noise variance of the lth measurement to the variance of the estimated value is defined as the interference rate,

[0174]

[0175] Definition 3: For a noisy measurement environment, when all contribution rates Less than the corresponding interference rate (V (1) ,V (2) ,...,V (m) ), the i-th state is weakly observable.

[0176] The third observability quantitative indicator R3 is:

[0177]

[0178] In formula (16), Indicates the measured average interference rate.

[0179] If the system is weakly observable, the interference rate is closer to 1 than the contribution rate, and the average interference rate is also close to 1. At this time, R3 is small and close to 0. Otherwise, R3 is large. At this time, the third indicator R3 can also reflect the degree of observability of the system through its numerical value.

[0180] (4) The impact of observation data loss on system observability

[0181] is the estimated variance of the lth measurement when no data is missing, is the variance of the estimated value of the lth measurement after data loss occurs. It represents the variance of the estimated value of the lth measurement due to missing observation data.

[0182] The ratio of the difference in variance before and after the observed data loss of the lth measurement to the variance of the estimated value when no data loss occurs is defined as the data loss impact factor.

[0183]

[0184] Definition 4: For a measurement scenario with data loss, when all contribution rates Less than the corresponding data loss impact factor (P (1) ,P (2) ,...,P (m) ), the i-th state is weakly observable.

[0185] The fourth observability quantitative indicator R4 is:

[0186]

[0187] In formula (18), Represents the average data loss impact factor of the measurement.

[0188] If the system is weakly observable, the data loss impact factor is closer to 1 than the contribution rate, and the average data loss impact factor is also close to 1. At this time, R4 is small and close to 0. On the contrary, R4 is large. At this time, the fourth indicator R4 can also reflect the degree of observability of the system through its numerical value.

[0189] (5) Impact of observation data delay on system observability

[0190] is the estimated value variance of the lth measurement without data delay, is the variance of the estimated value after the data delay occurs. It represents the variance of the estimated value of the lth measurement due to the delay of the observed data.

[0191] The ratio of the difference between the variances of the observed data before and after the data delay and the variance of the estimated value without data delay is defined as the data delay impact factor.

[0192]

[0193] Definition 5: For a measurement scenario with data delay, when all contribution rates Less than the corresponding data delay impact factor (T (1) ,T (2) ,...,T (m) ), the i-th state is weakly observable.

[0194] The fifth observability quantitative indicator R5 is:

[0195]

[0196] In formula (20), Indicates the measured average data delay impact factor.

[0197] If the system is weakly observable, the data delay impact factor is closer to 1 than the contribution rate, and the average data delay impact factor is also close to 1. At this time, R5 is small and close to 0. On the contrary, R5 is large. At this time, the fifth indicator R5 can also reflect the degree of observability of the system through its numerical value.

[0198] Combining the above five indicators, a comprehensive indicator R is proposed to quantify the impact of observability. total ,

[0199]

[0200] R total The value of reflects the strength of the system observability. The closer it is to 1, the stronger the system observability is; the closer it is to 0, the weaker the system observability is.

[0201] 4. Comprehensive analysis and quantitative indicators of system observability based on random response surface method

[0202] Let R o represents the rank of the observability coefficient matrix. According to the observability analysis method based on the random response surface method, if and only if the observability coefficient matrix Φ k With n linearly independent column vectors, the system is observable, that is, R o =n, the system is observable. The observability of a system is the premise for analyzing the strength of its observability.

[0203] The observability condition R o Combined with observability, a comprehensive index R that quantifies the degree of observability impact total , proposes an overall index R that can quantify the degree of observability of the system all .

[0204] R all =R o +R total ,R all ∈(n,n+1) (22)

[0205] According to the value of random interference rate, the total quantifiable system observability index R all It is divided into the following three levels to indicate the strength of observability.

[0206] When R all ∈(n,n+0.35), the system is weakly observable; when R all ∈(n+0.35,n+0.75), the observability of the system is moderate; when R all ∈(n+0.75,n+1), the system is strongly observable; according to R allThe size of can be divided into three levels of observability intensity: strong, medium, and weak;

[0207] If the observability condition R o <n, the system is unobservable. At this point, there is no need to further discuss the strength of the observability of the system. The above process is as follows Figure 1 shown.

[0208] Based on the above method design, the present invention also provides a static distribution network system observability evaluation device based on random response surface method, including an observability judgment module, which judges whether the distribution network system is observable on the basis of observability analysis of the static distribution network system using random response surface method; the basis is: if and only if the observability coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

[0209] Furthermore, it also includes a quantitative evaluation module: based on the observability judgment module, the observability of the static distribution network system is judged to be observable, and one or more of the following indicators are used to quantify the strength of the observability of the distribution network system, namely: the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system. For the selection of one, multiple or even all of them, those skilled in the art can refer to the description of the above method part.

[0210] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0211] It needs to be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, on which a computer program is stored, and the computer program is executed by a processor to execute the above method. The storage medium can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0212] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0213] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure may have various changes and improvements, and these changes and improvements fall within the scope of the present disclosure to be protected.

[0214] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of a static distribution network system observability evaluation method based on the random response surface method under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A static distribution network system observability evaluation method based on stochastic response surface method, characterized by: Based on the observability analysis of the static distribution network system using the random response surface method, the basis for judging whether the distribution network system is observable is: if and only if the observability coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

2. The method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 1 is characterized in that: The observability coefficient matrix is ​​obtained as follows: The random response surface method uses a truncated expanded surrogate model: The random output y is: Where φ0, φ1(ξ i ), Respectively represent the zero-order, first-order, and second-order random response surface method basis, a0, a i , a i,i represents the corresponding random response surface method coefficient; Zero-order random response surface method basis and its corresponding coefficients φ0 and represents the average of the estimated values ​​of the lth measurement, as well as the first-order and second-order random response surface method bases and their corresponding coefficients and represents the uncertainty of the estimated value of the lth measurement relative to the i-th state estimate, where the random response surface method coefficient and represents the contribution of the i-th state to the uncertainty of the estimated value of the l-th measurement; set up represents the observability coefficient matrix, which is a submatrix of the coefficient matrix with n linearly independent columns; as follows:

3. The method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 3 is characterized in that: On the basis of judging whether the static distribution network system is observable, the strength of the observability of the distribution network system is quantified by one or more of the following indicators, namely: the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system.

4. The method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 3 is characterized in that: The specific method to quantify the strength of the observability of the distribution network system is: Constructing a comprehensive index R to quantify the impact of observability total Among them, R1~R5 are the observability quantitative indicators corresponding to the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system; R total The value of reflects the strength of the system observability. The closer it is to 1, the stronger the system observability is; the closer it is to 0, the weaker the system observability is. Let R o represents the rank of the observability coefficient matrix; R o =n, the system is observable, which is the premise for analyzing the observability of the system; The observability condition R o Combined with observability, a comprehensive index R that quantifies the degree of observability impact total , construct an index R that can quantify the degree of observability of the system all ; R all =R o +R total ,R all ∈(n,n+1) When R all ∈(n,n+0.35), the system is weakly observable; when R all ∈(n+0.35,n+0.75), the observability of the system is moderate; when R all ∈(n+0.75,n+1), the system is strongly observable; If the observability condition R o <n, the system is unobservable.

5. A static distribution network system observability evaluation method based on stochastic response surface method according to claim 3 or 4, characterized in that: The strength of the observability of the distribution network system is quantified by the conditional quantification of the observability coefficient matrix: The condition number of the matrix is ​​the ratio of the largest singular value to the smallest singular value, that is: The reciprocal of the condition number is used as the observability quantitative indicator R1: The closer R1 is to 0, the weakly observable system is; the closer R1 is to 1, the observable system is.

6. A method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 3 or 4, characterized in that: The strength of the observability of the distribution network system is quantified by the contribution of the state to the estimation: According to the proxy model, the variance of the estimated value of the lth measurement is: The ratio of the contribution of the i-th state to the variance of the estimated value of the l-th measurement is defined as the contribution ratio, in, The observability quantitative index R2 is constructed as: If the system is strongly observable, R2 is close to 1; if the system is weakly observable, R2 is close to 0.

7. A method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 3 or 4, characterized in that: The observability of the distribution network system is quantified by observing the impact of noise on the observability of the system: The ratio of the noise variance of the lth measurement to the variance of the estimated value is defined as the interference rate: The observability quantitative index R3 is constructed as: in, Indicates the average interference rate of the measurement; If the system is weakly observable, R3 is close to 0.

8. A static distribution network system observability evaluation method based on stochastic response surface method according to claim 3 or 4, characterized in that: The influence of observation data loss on the observability of the system is used to quantify the strength of the observability of the distribution network system: is the estimated variance of the lth measurement when no data is missing, is the variance of the estimated value after the data loss occurs; represents the variance of the estimated value of the lth measurement due to missing observation data; The ratio of the difference in variance before and after the observed data loss of the lth measurement to the variance of the estimated value when no data loss occurs is defined as the data loss impact factor: The observability quantitative index R4 is constructed as: in, Indicates the average data loss impact factor of the measurement; If the system is weakly observable, R4 is close to 0.

9. A method for evaluating the observability of a static distribution network system based on a random response surface method according to claim 3 or 4, characterized in that: The influence of observation data delay on system observability is used to quantify the strength of the distribution network system observability: is the estimated value variance of the lth measurement without data delay, is the variance of the estimated value after the data delay occurs; represents the variance of the estimated value of the lth measurement due to the delay of the observed data; The ratio of the difference between the variances of the observed data before and after the data delay and the variance of the estimated value without data delay is defined as the data delay impact factor. The observability quantitative index R5 is constructed as follows: in, Indicates the measured average data delay impact factor; If the system is weakly observable, R5 is close to 0.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, the method described in any one of claims 1 to 9 is adopted.

11. A static distribution network system observability evaluation device based on random response surface method, characterized in that: It includes an observability judgment module, which judges whether the distribution network system is observable based on the observability analysis of the static distribution network system using the random response surface method; the basis is: if and only if the observability coefficient matrix Φ k The system is observable when there are n linearly independent column vectors, and the i-th row and the n+i-th row cannot be zero vectors at the same time.

12. The static distribution network system observability assessment device based on stochastic response surface method according to claim 11, characterized in that: It also includes a quantitative evaluation module: based on the observability judgment module, it is used to quantify the strength of the observability of the distribution network system through one or more of the following indicators, namely: the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system.

13. The static distribution network system observability assessment device based on stochastic response surface method according to claim 11, characterized in that: The quantitative evaluation module uses the comprehensive index R total Among them, R1~R5 are the observability quantitative indicators corresponding to the condition number of the observability coefficient matrix, the contribution of the state to the estimation, the impact of observation noise on the observability of the system, the impact of observation data loss on the observability of the system, and the impact of observation data delay on the observability of the system; R total The value of reflects the strength of the system observability. The closer it is to 1, the stronger the system observability is; the closer it is to 0, the weaker the system observability is. Let R o represents the rank of the observability coefficient matrix; R o =n, the system is observable, which is the premise for analyzing the observability of the system; The observability condition R o Combined with observability, a comprehensive index R that quantifies the degree of observability impact total , construct an index R that can quantify the degree of observability of the system all ; R all =R o +R total ,R all ∈(n,n+1) When R all ∈(n,n+0.35), the system is weakly observable; when R all ∈(n+0.35,n+0.75), the observability of the system is moderate; when R all ∈(n+0.75,n+1), the system is strongly observable; If the observability condition R o <n, the system is unobservable.

Images