CN-119716967-B - Recursion method for automatically identifying earthquake event on line

Abstract

The invention discloses a recursion method for automatically identifying earthquake events on line, which is used for reconstructing a phase space of a time-course signal of earthquake vibration, further visually presenting the reconstructed phase space by adopting a recursion diagram, quantitatively analyzing recursion diagram information, and determining an alarm threshold value when the earthquake occurs, thereby identifying the earthquake events according to recursion quantitative indexes. According to the invention, as the recursion diagram and the improved recursion quantization index are adopted, the problems of non-intuitiveness and low recognition timeliness of the conventional earthquake event recognition result are solved, the recursion diagram can graphically display the earthquake event, and the improved recursion quantization index can realize real-time online early warning of the earthquake event; the method can be used for monitoring and alarming the bridge earthquake events in real time, and improves intuitiveness and recognition timeliness of the bridge earthquake event recognition result.

Inventors

• LIANG PENG
• Lu Xingshun
• ZHANG DI
• HE MIN

Assignees

• 长安大学

Dates

Publication Date

20260512

Application Date

20241212

Claims (2)

1. 1. A recursive method for on-line automatic identification of seismic events, comprising the steps of: step 1, reconstructing a phase space, namely performing phase space reconstruction on a seismic event time sequence by adopting a delay coordinate method, and determining an optimal embedding dimension and an optimal delay time required by the phase space reconstruction; Step 2, drawing a recursion chart, namely drawing the recursion chart of the phase space reconstruction result, and determining parameters required by drawing the recursion chart; step 3, determining a recursion quantization index, namely performing quantization analysis on the recursion graph to obtain a seismic identification quantization index; step 4, early warning is carried out on the earthquake event according to the distribution rule of the earthquake identification quantization index; In the step 1, the phase space reconstruction is performed on the time sequence as follows: In the formula, For the time series matrix after phase space reconstruction, Representing the time; Represent the first A time sequence of moments; In order to embed the dimensions of the dimensions, Is a delay time; The method comprises determining the embedding dimension by false neighboring point method Defining a time sequence Is that Solving a series of phase point vectors in the dimensional-phase space, solving false adjacent points, calculating the proportion of the false adjacent points corresponding to different embedding dimensions to determine the optimal embedding dimension, wherein the proportion of the false adjacent points is less than 5% or the false adjacent points are no longer associated with the false adjacent points When the number of the embedded dimension is increased and decreased, the corresponding dimension is the optimal embedded dimension; In the step 1, the determining process of the optimal delay time includes: determination of delay time by mutual information method Defining a time sequence And delay time Time series of (2) For two groups of discrete sequences, according to the principle of information theory, the mutual information of the two groups of sequences is obtained If (1) Then And (3) with Uncorrelated; The minimum value of (2) represents And (3) with Is the largest possible uncorrelation; the first minimum point in (2) is the optimal delay time; In the step 2, the recursion chart is a graph formed by a 0-1 matrix and two time axes which are described by different colors, and the definition formula of the distance between any two vectors is as follows: wherein: Is one A two-dimensional matrix of steps, Represent the first At the moment of time of day, Is a state vector Is used in the number of (a) and (b), Is the dimension of the embedding, It is the delay time that is set to be, Is the cut-off threshold value, Representing the euclidean norm, Representation of The function of the function is that, In the recursion diagram, different The values are represented in two easily distinguishable colors; in the step3, the earthquake identification quantization index : Wherein N represents the sampling point number of the recursion chart, TT is the average vertical line in the recursion chart, and RR is the recursion rate; the value range is [0, N/2], The value range is 0, 1; The cut-off threshold The selection method of (2) is a dynamic threshold method, and the selection standard is 20% of the maximum phase space diameter ; When (when) All points are recursive points, and the recursive graph is pure black when When the signal is divided into a recursive point and a non-recursive point, the recursive graph is formed by intersecting black and white lines; The TT calculation formula is as follows: Where N represents the number of sampling points of the recursion graph, The length of the vertical line in the recursion diagram is shown, Representing the length of the minimum vertical line, Is the length of the recursion diagram Is a distribution probability of vertical lines of (a); in the step 4, when the earthquake is identified and quantified Distributed near 0, then being random signal, when earthquake is identified and quantified And the distribution range is (0.8,1) which is towards 1, and then the earthquake signal is used for earthquake early warning.
2. 2. The recursive method for online automatic identification of a seismic event of claim 1, wherein the RR calculation formula is as follows: Where N represents the number of sampling points of the recursion graph, Represent the first At the moment of time of day, Is a recursive value.

Description

Recursion method for automatically identifying earthquake event on line Technical Field The invention belongs to the field of bridge health monitoring earthquake event identification and early warning, and relates to an automatic earthquake event identification and early warning method. Background The real-time online monitoring and early warning of the earthquake event is an important content of bridge health monitoring, and the existing research method mainly comprises a long-short time window ratio method, a red pool information criterion, a comprehensive analysis algorithm of machine learning and various methods. The existing method has the problems of poor real-time performance, low automation degree and the like of seismic event processing, and has high difficulty in real-time monitoring and alarming of the seismic event, and decision suggestions can not be timely provided for bridge maintenance. Disclosure of Invention Aiming at the defects existing in the prior art, the invention aims to provide a recursion method for automatically identifying the earthquake event on line, which introduces a recursion diagram and improved recursion quantization indexes to solve the problems of low accuracy, non-visual result and low identification timeliness of the earthquake event, so as to realize graphical display and real-time early warning of the earthquake event. In order to solve the technical problems, the invention adopts the following technical scheme: a recursive method for online automatic identification of seismic events, comprising the steps of: step 1, reconstructing a phase space, namely performing phase space reconstruction on a seismic event time sequence by adopting a delay coordinate method, and determining an optimal embedding dimension and an optimal delay time required by the phase space reconstruction; Step 2, drawing a recursion chart, namely drawing the recursion chart of the phase space reconstruction result, and determining parameters required by drawing the recursion chart; step 3, determining a recursion quantization index, namely performing quantization analysis on the recursion graph to obtain a seismic identification quantization index; And 4, early warning is carried out on the earthquake event according to the distribution rule of the earthquake identification quantization index. The invention also comprises the following technical characteristics: specifically, in the step 1, the phase space reconstruction is performed on the time sequence according to the following formula: X(i)={x(i),x(i+τ),...,x(i+(m-1)τ)} Wherein X (i) is a time sequence matrix after phase space reconstruction, i represents time, X (i+ (m-1) τ) represents time sequence of i+ (m-1) τ, m is embedding dimension, and τ is delay time. Specifically, in the step 1, the process of determining the optimal embedding dimension includes adopting a false adjacent point method to determine the embedding dimension m, defining a time sequence x (t) as a series of phase point vectors in m-dimensional phase space, solving false adjacent points, calculating the proportion of false adjacent points corresponding to different embedding dimensions to determine the optimal embedding dimension, and when the proportion of the false adjacent points is less than 5% or the false adjacent points are not reduced along with the increase of m, determining the corresponding dimension as the optimal embedding dimension. Specifically, in the step 1, the determining process of the optimal delay time includes determining the delay time τ by adopting a mutual information method, wherein the time sequence x (t) and the time sequence x (t+τ) of the delay time τ are defined as two groups of discrete sequences, the mutual information I (τ) of the two groups of sequences is obtained according to an information theory principle, if I (τ) =0, the x (t+τ) is uncorrelated with the x (t), the minimum value of the I (τ) indicates that the x (t+τ) and the x (t) are the largest possible uncorrelation, and the first minimum value point in the I (τ) is the optimal delay time. Specifically, in the step 2, the recursive graph is a graph formed by a 0-1 matrix described by different colors and two time axes, and the definition formula of the distance between any two vectors is as follows: Wherein R i,j is a two-dimensional matrix of N- (m-1) τ order, i, j represents the i, j th moment, N is a state vector M is the embedding dimension, τ is the delay time, ε is the cutoff threshold, |·|| represents the euclidean norm, Θ (·) represents the Heaviside function, Θ (x+.0) =0, Θ (x > 0) =1, and in the recursive graph the different R i,j values are represented by two easily distinguishable colors. Specifically, the selection method of the cut-off threshold epsilon is a dynamic threshold method, and the selection standard is 20% of the maximum phase space diameter r m; when epsilon is larger than or equal to r m, all points are recursive points, the recursive graph is pure b