CN107909757B - Vibration source early warning method of optical fiber vibration safety early warning system based on sequential detection - Google Patents

Abstract

The invention discloses a vibration source early warning method of an optical fiber vibration safety early warning system based on sequential detection, which comprises the following steps: acquiring optical fiber vibration signal data, determining the spatial position of the vibration signal data in an optical fiber, and establishing a two-dimensional array based on time and spatial arrangement, wherein: the abscissa of the two-dimensional array is the spatial position axis, the ordinate of the two-dimensional array is a time axis, and the early warning method further comprises the following steps: the first step is as follows: establishing a two-dimensional sequential detection model according to the two-dimensional array; the second step is that: determining an alarm level through the established two-dimensional sequential detection model, and executing alarm output; compared with the prior art, the method has the advantages of high response speed, low false alarm rate and low false alarm rate, and adopts a sequential method to carry out cumulative analysis on the alarm degree, thereby ensuring the contribution of each vibration to the alarm degree and realizing the quick response to the continuous vibration source.

Description

Vibration source early warning method of optical fiber vibration safety early warning system based on sequential detection

technical field

The invention relates to a vibration source early-warning method for an optical fiber vibration safety early-warning system based on sequential detection. solution.

Background technique

With the rapid development of urban construction, more and more building facilities and areas need to be protected, such as military facilities, power stations, hospitals, oil pipelines, etc. The optical fiber vibration safety early warning system adopts passive distributed optical fiber as the sensor, which has the advantages of strong anti-interference ability and large detection distance. The optical fiber vibration safety warning system detects the external vibration signals along the optical fiber, and automatically identifies the type of vibration source through the relevant processing of the detection signal, so as to realize the timely alarm of harmful behaviors, which has important practical value.

At present, the signal detection technology is relatively mature. The signal-to-noise ratio analysis is performed on the vibration data collected by the sensor, and the signal part is extracted to determine the presence or absence of the signal. However, it is difficult to identify the signal type of the detected signal. At present, the commonly used vibration source identification methods include linear classifiers, neural networks, variance analysis, chaos analysis, support vector machines, etc. These methods all need to learn and train a large number of samples, require a large amount of calculation, and have problems such as training convergence.

For the vibration signal detected in the early warning system, once the vibration source type is determined, the corresponding early warning processing can be carried out. However, in most cases, the vibration source type has the characteristics of difficult identification and many new types of vibration sources. Therefore, for In this case, it is necessary to analyze the continuity of the vibration source to reduce the system missed alarm rate and false alarm rate. Aiming at the false alarm rate and missed alarm rate in the optical fiber security early warning system.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a vibration source early warning method for an optical fiber vibration safety early warning system based on sequential detection, analyze the possibility of harmful vibration sources at the same position (that is, the alarm degree), and establish a second-level early warning strategy model, reducing the The false alarm rate and missed alarm rate of the system, and at the same time, the automatic classification of alarm types is realized, which enhances the practical value of the system.

In order to achieve the above object, the technical scheme of the present invention is: a vibration source early warning method for an optical fiber vibration safety early warning system based on sequential detection, comprising: acquiring optical fiber vibration signal data, determining the spatial position of the vibration signal data in the optical fiber, establishing a time-based and the two-dimensional array arranged in space, wherein: the abscissa of the two-dimensional array is the space position axis, the ordinate of the two-dimensional array is the time axis, and the early warning method step further comprises:

The first step: establish a two-dimensional sequential detection model based on a two-dimensional array;

Step 2: Determine the alarm level through the established two-dimensional sequential detection model, and execute the alarm output;

The two-dimensional sequential detection model is the formula:

formula:

L(t,k) represents the sequential alarm degree at the kth position at time t;

ΔL(t,k) represents the sequential alarm degree increment at the kth position at time t;

t represents time, on the time axis, the unit is seconds;

k represents the position point of the vibration signal on the spatial position axis;

k _n is the length of the protection interval of the nth point in the spatial position axis direction;

x _t,j represents the vibration signal amplitude coefficient at time t, j=kk _n position;

S(x _t,j ) represents the alarm degree coefficient when the vibration signal amplitude coefficient is x _t,j ;

α is the increasing factor of sequential alarm degree, α>0;

γ is the sequential alarm reduction factor, γ<0;

ε _k is the vibration signal continuity factor at the k-th position;

Described through the established two-dimensional sequential detection model to determine the alarm level, the execution alarm output is:

Set the alarm level to two levels, which are the first-level warning and the second-level alarm, respectively, and set the warning threshold and alarm threshold respectively;

Use the two-dimensional sequential detection model to calculate the two-dimensional array to obtain the sequential alarm degree of the vibration signal data in the spatial position of the optical fiber, (all vibration signals after preprocessing and alarm degree mapping are sequentially alarmed. degree calculation;) compare the sequential alarm degree with the warning threshold and the alarm threshold:

When the sequential alarm degree is higher than the early warning threshold, it is determined that the early warning alarm signal is issued, and the warning time and location information are recorded;

When the sequential alarm degree is higher than the alarm threshold, it is determined that the advanced alarm signal is issued for the advanced alarm, and the alarm time and location information are recorded;

When the sequential alarm degree continuously exceeds a set duration threshold between the early warning threshold and the alarm threshold, it is determined that a low-level alarm is sent out a low-level alarm signal, and the alarm time and location information are recorded.

The scheme is further: the determining the spatial position of the vibration signal data in the optical fiber is: dividing the acquired optical fiber vibration signal data into the section position of the vibration signal in the optical fiber through spectrum analysis, and the section position is the spatial position of the optical fiber. .

The scheme is further: the method steps also include:

When the sequential alarm degree at the same position decreases from higher than the warning threshold to lower than the warning threshold, the warning signal is released, and the release time and location information are recorded.

The solution is further: the acquisition formula of the vibration signal continuity factor is:

Among them: t _n is the protection time length of sequential alarm degree drop.

The solution is further: the method steps further include: establishing an alarm level confirmation model, and automatically marking the alarm level, and the alarm level confirmation model can be expressed as a formula:

n _k =f(t _2,k -t _1,k )

in:

n _k represents the alarm level of the kth position point;

t _1,k represents the warning time of the kth position point;

t _2,k represents the alarm time of the kth position point.

The solution is further: the value range of the alarm degree coefficient is 0-1, the sequential alarm degree increasing factor is equal to 1, and the sequential alarm degree decreasing factor is equal to -0.2.

Compared with the prior art, the beneficial effects of the present invention:

1. The response speed is fast, and the sequential method is used to accumulate and analyze the alarm degree to ensure the contribution of each vibration to the alarm degree, and to achieve a rapid response to the continuous vibration source;

2. The false alarm rate is low, the second-level early warning strategy model is adopted, and the vibration data to the alarm degree mapping function model is used to avoid the influence of the interference signal on the alarm strategy and reduce the false alarm rate of the system;

3. The leakage alarm rate is low, and an algorithm based on the sequential alarm degree is established. It has a certain memory function for intermittent signals and achieves a low and low alarm leakage rate.

The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

Description of drawings

Fig. 1 is a schematic diagram of the preprocessed field vibration signal data of the present invention;

2 is a schematic diagram of the two-dimensional sequential alarm degree algorithm of the present invention;

Figure 3 is a schematic diagram of the calculation result of the sequential alarm degree;

FIG. 4 is a schematic diagram of the principle of the two-level early warning strategy of the present invention.

Detailed ways

A vibration source early warning method for an optical fiber vibration safety early warning system based on sequential detection, comprising: acquiring optical fiber vibration signal data, preprocessing the acquired vibration signal data, and determining the spatial position of the vibration signal data in the optical fiber, as shown in Figure 1 , establish a two-dimensional array arranged based on time and space, wherein: the abscissa of the two-dimensional array is the space position axis, the ordinate of the two-dimensional array is the time axis, and the early warning method steps further include:

The first step: establish a two-dimensional sequential detection model based on a two-dimensional array;

Step 2: Determine the alarm level through the established two-dimensional sequential detection model, and execute the alarm output;

The two-dimensional sequential detection model is the formula:

formula:

L(t,k) represents the sequential alarm degree at the kth position at time t;

ΔL(t,k) represents the sequential alarm degree increment at the kth position at time t;

t represents time, on the time axis, the unit is seconds;

k represents the position point of the vibration signal on the spatial position axis;

k _n is the length of the protection interval of the nth point in the spatial position axis direction;

x _t,j represents the vibration signal amplitude coefficient at time t, j=kk _n position; x _t,j ≥ 0

S(x _t,j ) represents the alarm degree coefficient when the vibration signal amplitude coefficient is x _t,j ;

α is the increasing factor of sequential alarm degree, α>0;

γ is the sequential alarm reduction factor, γ<0;

ε _k is the vibration signal continuity factor at the k-th position;

Figure 2 shows the principle of the two-dimensional sequential alarm degree algorithm, which is the embodiment of Figure 1. The horizontal direction represents the segment position of the optical fiber, and Figure 3 is the corresponding sequential alarm degree calculation result curve.

Described through the established two-dimensional sequential detection model to determine the alarm level, the execution alarm output is:

As shown in Figure 4, the alarm level is set to two levels, which are the first-level warning and the second-level alarm, respectively, and the warning threshold threshold, as shown in Figure 4, and the alarm threshold threshold, as shown in Figure 4, are set respectively. high threshold;

Use the two-dimensional sequential detection model to calculate the two-dimensional array to obtain the sequential alarm degree of the vibration signal data in the spatial position of the optical fiber, (all vibration signals after preprocessing and alarm degree mapping are sequentially alarmed. degree calculation;) compare the sequential alarm degree with the warning threshold and the alarm threshold:

When the sequential alarm degree is higher than the early warning threshold, it is determined that the early warning alarm signal is issued, and the warning time and location information are recorded;

When the sequential alarm degree is higher than the alarm threshold, it is determined that the advanced alarm signal is issued for the advanced alarm, and the alarm time and location information are recorded;

When the sequential alarm degree continuously exceeds a set duration threshold between the early warning threshold and the alarm threshold, it is determined that a low-level alarm is sent out a low-level alarm signal, and the alarm time and location information are recorded; the time threshold can be set according to actual needs. For example, it can be hundreds of milliseconds, or several seconds or tens of seconds.

Wherein: the determining the spatial position of the vibration signal data in the optical fiber is: dividing the acquired optical fiber vibration signal data into the section position of the vibration signal in the optical fiber through spectrum analysis, and the section position is the spatial position of the optical fiber.

The preprocessing signal is a real number not less than 0. The larger the value, the stronger the vibration and the greater the contribution to the alarm degree. The vibration signal data is actually the optical fiber vibration spectrum signal. It is a well-known and mature technology to find out the position of the optical fiber segment corresponding to the spectral signal, and then determine the spatial position of the vibration signal data in the optical fiber.

In an embodiment: the method steps further include:

When the sequential alarm degree at the same position decreases from higher than the warning threshold to lower than the warning threshold, the warning signal is released, and the release time and location information are recorded.

Wherein: the formula for obtaining the continuity factor of the vibration signal is:

Among them: t _n is the protection time length of sequential alarm degree drop.

Wherein: the method steps further include: establishing an alarm level confirmation model, and automatically marking the alarm level, and the alarm level confirmation model can be expressed as a formula:

n _k =f(t _2,k -t _1,k )

in:

n _k represents the alarm level of the kth position point;

t _1,k represents the warning time of the kth position point;

t _2,k represents the alarm time of the kth position point.

The preferred solution is: the value range of the alarm degree coefficient is 0-1, the sequential alarm degree increasing factor α is equal to 1, and the sequential alarm degree decreasing factor γ is equal to -0.2.

In the embodiment, the preprocessed signal is a number not less than 0, and the data is greater than zero only when the presence of the vibration signal is detected. Due to the large amount of preprocessed signal data, in order to avoid a large preprocessed data causing system alarm, the alarm degree coefficient S(x _t,j ) is set, and its value range is 0-1. The mapping relationship between the data and the alarm degree ensures that the influence of each vibration on the alarm degree varies between 0 and 1. The relationship is as follows:

in:

x _t,j represents the amplitude coefficient of the preprocessed vibration signal, x _t,j ≥ 0;

S(x _{t, j} ) represents the alarm degree when the signal amplitude coefficient is x _{t, j} , and the value ranges from 0 to 1;

β represents the change factor of the alarm degree with the vibration signal, β>0, it is an empirical parameter, usually β=1;

S ₀ represents the adjustment coefficient of alarm degree;

In order to prevent the deviation of the collected vibration signals due to the positioning accuracy, k _{n is} used as the length of the protection interval, as shown in the horizontal long dashed box in Figure 2. At the same time, considering the signal continuity factor, t _n is set as the signal continuity judgment threshold, as shown in the vertical dotted box in Figure 2, the corresponding vibration signal continuity judgment factor relationship is:

In the above embodiment, the schematic diagram of the two-dimensional sequential alarm degree algorithm is shown in Figure 2. The calculation of the sequential alarm degree at time t and position k involves the vibration data at time t _n and k _n points left and right. By calculating the sequential alarm degree information of all points, combined with the two-level early warning strategy model and the early warning level model shown in Figure 4, the early warning level analysis of the vibration source at the current location is carried out to realize the automatic processing of the alarm and the alarm level. Its advantages are: the response speed is fast, and the sequential method is used to analyze the alarm degree cumulatively, to ensure the contribution of each vibration to the alarm degree, and to achieve rapid response to the continuous vibration source; the false alarm rate is low, and the second-level early warning strategy is adopted. The model, combined with the vibration data to the alarm degree mapping function model, avoids the influence of the interference signal on the alarm strategy and reduces the false alarm rate of the system; the missed alarm rate is low, and an algorithm based on the sequential alarm degree is established, which has a certain memory for intermittent signals. function to achieve a lower low leakage alarm rate.

Claims (5)

1. A vibration source early warning method for an optical fiber vibration safety early warning system based on sequential detection, comprising: acquiring optical fiber vibration signal data, determining the spatial position of the vibration signal data in the optical fiber, and establishing a two-dimensional array based on time and space arrangement, wherein: The abscissa of the two-dimensional array is the spatial position axis, and the ordinate of the two-dimensional array is the time axis. It is characterized in that, the steps of the early warning method further include: The first step: establish a two-dimensional sequential detection model based on a two-dimensional array; Step 2: Determine the alarm level through the established two-dimensional sequential detection model, and execute the alarm output; The two-dimensional sequential detection model is the formula: formula: L(t,k) represents the sequential alarm degree at the kth position at time t; ΔL(t,k) represents the sequential alarm degree increment at the kth position at time t; t represents time, on the time axis, the unit is seconds; k represents the position point of the vibration signal on the spatial position axis; k _n is the length of the protection interval of the nth point in the spatial position axis direction; x _t,j represents the vibration signal amplitude coefficient at time t, j=kk _n position; S(x _t,j ) represents the alarm degree coefficient, and the value range is 0-1; α is the increasing factor of sequential alarm degree, α>0; γ is the sequential alarm reduction factor, γ<0; ε _k is the continuity factor, Among them: t _n is the protection time length of sequential alarm degree drop; Described through the established two-dimensional sequential detection model to determine the alarm level, the execution alarm output is: Set the alarm level to two levels, which are the first-level warning and the second-level alarm, respectively, and set the warning threshold and alarm threshold respectively; Use the two-dimensional sequential detection model to calculate the two-dimensional array, obtain the sequential alarm degree of the vibration signal data in the spatial position of the optical fiber, and compare the sequential alarm degree with the warning threshold and the alarm threshold: When the sequential alarm degree is higher than the early warning threshold, it is determined that the early warning alarm signal is issued, and the warning time and location information are recorded; When the sequential alarm degree is higher than the alarm threshold, it is determined that the advanced alarm signal is issued for the advanced alarm, and the alarm time and location information are recorded; When the sequential alarm degree continuously exceeds a set duration threshold between the early warning threshold and the alarm threshold, it is determined that a low-level alarm is sent out a low-level alarm signal, and the alarm time and location information are recorded. 2. The method according to claim 1, wherein the determining the spatial position of the vibration signal data in the optical fiber is: dividing the acquired optical fiber vibration signal data into the segment position of the vibration signal in the optical fiber through spectrum analysis, so The aforementioned segment position is the spatial position of the optical fiber. 3. The method according to claim 1, wherein the method step further comprises: When the sequential alarm degree at the same position decreases from higher than the warning threshold to lower than the warning threshold, the warning signal is released, and the release time and location information are recorded. 4. The method according to claim 1, wherein the method step further comprises: establishing an alarm level confirmation model, automatically labeling the alarm level, and the alarm level confirmation model can be expressed as a formula: n _k =f(t _2,k -t _1,k ) in: n _k represents the alarm level of the kth position point; t _1,k represents the warning time of the kth position point; t _2,k represents the alarm time of the kth position point. 5 . The method according to claim 1 , wherein the sequential alarm degree increasing factor is equal to 1, and the sequential alarm degree decreasing factor is equal to -0.2. 6 .