CN104373820B - The method for reducing line leakage rate of false alarm - Google Patents

Abstract

The invention discloses a method for reducing the false alarm rate of pipeline leakage monitoring. The method comprises the following steps: respectively real-time and continuously monitoring the acoustic wave signal of the first station and the acoustic wave signal of the last station inside the pipeline through the first acoustic wave monitor arranged at the first station of the pipeline and the second acoustic wave monitor at the last station of the pipeline; When there are abnormal signals at the first and last stations, calculate the peak value of the abnormal signal in the acoustic signal of the first station and the peak value of the abnormal signal in the acoustic signal of the last station; judge the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station Whether it is within the preset range, if yes, the abnormal signal of the first and last stations is a signal of the same source, and a leak alarm is issued; if not, the abnormal signal of the first and last station is a non-homologous signal, and a leak alarm is not issued. It can prevent the system from issuing false alarms when it is judged that the signal is not of the same source, can effectively reduce the false alarm rate of existing pipeline leakage monitoring, and improve the reliability of leakage monitoring, and this method has good robustness.

Description

Method for reducing the false alarm rate of pipeline leakage monitoring

technical field

The invention relates to the field of pipeline leakage monitoring, in particular to a method for reducing the false alarm rate of pipeline leakage monitoring.

Background technique

In the pipeline transportation of oil and gas, hazardous chemicals, etc., the pipeline laying distance is long and the line is complicated, and pipeline leakage due to corrosion and aging, engineering construction, and man-made damage often occurs, causing major leakage and explosion of flammable materials, environmental pollution, and casualties. loss.

At present, the main monitoring methods for pipeline leakage monitoring are: acoustic wave method, optical fiber method, negative pressure wave method, etc. Among them, the pipeline leakage monitoring method based on acoustic waves is low in cost, easy to implement, and has high leakage monitoring sensitivity and positioning accuracy, and has received a lot of attention in recent years. The types of sensors used in the acoustic wave method include: acoustic wave sensors, acceleration sensors, pickups, and piezoelectric pressure sensors. However, no matter which sensor is selected for pipeline leakage monitoring, false alarms will inevitably be caused due to the high sensitivity of the acoustic sensor, especially when the upstream and downstream monitoring stations monitor different interference signals respectively, it will inevitably cause false alarms.

Therefore, it is of great significance to explore a reliable identification method of abnormal signal homology based on the acoustic signal propagation attenuation model to reduce the false alarm rate and improve the reliability of the leakage monitoring system.

Contents of the invention

Based on this, in order to realize the object of the present invention, a method for reducing the false alarm rate of pipeline leakage monitoring comprises the following steps:

The acoustic wave signal of the first station and the acoustic wave signal of the last station inside the pipeline are respectively monitored in real time and continuously by the first acoustic wave monitor installed at the first station of the pipeline and the second acoustic wave monitor at the last station of the pipeline;

When abnormal signals are detected at both the first station and the last station, the peak value of the first station abnormal signal in the first station sound wave signal and the last station abnormal signal peak value in the last station sound wave signal are calculated respectively;

Judging whether the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is within a preset range, if yes, the pipeline leaks; if not, the pipeline does not leak.

As a possible implementation of the method for reducing the false alarm rate of pipeline leakage monitoring, the first acoustic wave monitor installed at the first station of the pipeline and the second acoustic wave monitor at the end station of the pipeline respectively monitor the interior of the pipeline in real time and continuously The sound wave signal of the first station and the sound wave signal of the last station include the following steps:

Sampling the two-way signals of the first-station sound wave signal and the last-station sound wave signal output by the first sound wave monitor and the second sound wave monitor respectively with a sampling period T, and the collected two-way The signals are time-stamped respectively;

respectively extracting the acoustic wave signal of the first station and the acoustic signal of the last station for two consecutive minutes, wherein the signal of the previous minute is historical data, and the signal of the next minute is real-time data, and the historical data and the real-time data constitute a complete frame The first-station acoustic signal and the last-station acoustic signal to be processed;

Wherein, T is a positive number greater than 0.

As an implementable method of reducing the false alarm rate of pipeline leakage monitoring, it also includes the step of determining the preset range of the ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station. This step specifically includes the following steps :

Obtain multiple sets of first-station acoustic signals and last-station acoustic signals and perform filtering, denoising and de-averaging processing to obtain denoised signals;

The peak position of the correlation coefficient of the abnormal signal at the first station and the abnormal signal at the last station is obtained by correlation calculation on the denoised signal, which is denoted as CorrPos, and according to the formula:

Determine the distance from the location of the abnormal signal to the first station, where l is the distance from the location of the abnormal signal to the first station, L is the total length of the pipeline, V _dn and V _up are from the first station to the last station and from the last station to the first station, respectively. Acoustic signal propagation speed of the station;

After monitoring for a preset time, the abnormal signal of the first station with the positioning position as the first station and the abnormal signal of the last station with the same source, as well as the abnormal signal of the first station with the positioning position as the last station and the abnormal signal of the last station with the same source are obtained, and the second A preset number of normal signals of the first station and normal signals of the last station;

performing discrete Fourier transform on the denoised signal to obtain the frequency spectrum of the denoised signal;

comparing the frequency spectrum of the normal signal and the abnormal signal, and intercepting the frequency spectrum of the abnormal signal;

performing an inverse Fourier transform on the intercepted frequency spectrum of the abnormal signal to obtain a time-domain reconstructed signal of the abnormal signal;

Determine the peak value ratio of the peak value of the abnormal signal at the first station to the peak value of the abnormal signal at the end station of the same source according to the amplitude of the reconstructed signal in the time domain;

According to the propagation attenuation formula of sound wave along the pipeline:

Among them, Peak ₀ is the initial amplitude of the sound wave signal at the point where the abnormal sound wave occurs, Peak is the signal amplitude after propagation attenuation, l is the distance from the location of the abnormal signal to the first station, L is the total length of the pipeline, and α is the sound wave propagation attenuation coefficient , when the acoustic signal generated by the first station propagates downstream to the last station, and when the acoustic signal generated by the last station propagates countercurrently to the first station, they respectively satisfy:

Peak _up is the amplitude of the abnormal signal collected by the first station, Peak _dn is the amplitude of the abnormal signal of the same source collected by the last station, α _s and α _n are the propagation attenuation coefficients when propagating downstream and upstream respectively;

Further calculation can obtain the attenuation coefficient α _s and the attenuation coefficient α _n of the acoustic wave signal in the downstream propagation as follows:

Obtain the values of α _s and α _n according to the peak ratio of the abnormal signal of the first station and the homologous abnormal signal of the last station;

Combined with the peak ratio of the second preset number of pairs of abnormal signals at the first station and the abnormal signals at the end station of the same source, the ranges of the attenuation coefficient α _s for the downstream propagation of the acoustic wave signal and the attenuation coefficient α _n for the upstream propagation of the acoustic wave signal are:

α _{smin ≤} α _s ≤ α _smax

α _{nmin ≤} α _{n ≤} α _nmax ;

For the leakage acoustic wave signal that occurs at a distance l from the first station, the peak ratio r of the abnormal signal at the first station and the abnormal signal at the end station of the same source satisfies the following relationship:

The sound wave propagation attenuation coefficients are divided into four groups [α _smin ,α _nmin ],[α _smin ,α _nmax ],[α _smax ,α _nmin ],[α _smax ,α _nmax ], and four r values are obtained, and the maximum It is recorded as r _max and the minimum value is recorded as r _min , then the value range of the peak ratio of the abnormal signal at the first station and the abnormal signal at the end station of the same source is:

As an implementable method for reducing the false alarm rate of pipeline leakage monitoring, it is judged whether the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is within a preset range, and if so, the pipeline leakage Leak; if no, then the pipeline does not leak, including the following steps:

calculating the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station;

According to the formula and the value of α _s and α _n to calculate the preset range of the peak ratio;

judging whether the peak ratio is within the calculated preset range;

If so, the abnormal signal of the first and last station is a signal of the same source, and a leak alarm is issued;

If not, the abnormal signal of the first and last station is a non-homologous signal, and no leakage alarm will be issued.

As a method for reducing the false alarm rate of pipeline leakage monitoring, the determination of the peak value ratio of the peak value of the abnormal signal at the first station to the peak value of the abnormal signal at the last station according to the amplitude of the reconstructed signal in the time domain includes the following step:

According to the polarity of the amplitude, the positive and negative intervals are divided, and the peak value of each interval is calculated as Peak[i], the maximum value is taken in the positive interval, and the minimum value is taken in the negative interval, wherein, 1≤i≤N _Peak , N _Peak is the total number of intervals in a complete frame of the first-station acoustic signal or the last-station acoustic signal to be processed;

Calculate the mean value of the positive peak value and the negative peak value, which are respectively recorded as meanVP and meanVN, and calculate the normalized peak prominence index L _P for the peak value greater than meanVP in the positive signal peak value according to the following formula, and when the positive peak value is smaller than meanVP, let L _P be 0,

Calculate the normalized peak prominence index L _N of the negative peak with an amplitude greater than meanVN according to the following formula, and set L _N to 0 when the negative peak is greater than meanVN,

Clearing the signals of the signal intervals in which the normalized peak prominence index is lower than the preset threshold to 0, and determining the peak interval and occurrence time LeakPos of the normalized peak prominence index higher than the preset threshold;

When there are multiple abnormal signals, calculate the LeakPos difference between the abnormal signal at the first station and the abnormal signals at the last station in turn:

DT＝LeakPos _up (i)-LeakPos _dn (j)

When the difference DT is the closest to the CorrPos, it is determined that this group of signals is respectively the abnormal signal of the first station and the abnormal signal of the corresponding last station. At this time, the interval sequence numbers are respectively recorded as Range _up and Range _dn ;

Further, the peak value Peak(Range _up ) of the abnormal signal at the first station and the peak value Peak(Range _dn ) of the abnormal signal at the last station are obtained, and the peak value ratio of the abnormal signal at the first station to the peak value of the abnormal signal at the last station is further obtained.

As a method for reducing the false alarm rate of pipeline leakage monitoring, the first acoustic wave monitor at the first station of the pipeline and the second acoustic wave monitor at the last station of the pipeline are used to monitor the acoustic signal of the first station and the sound wave of the last station inside the pipeline Signal time, using the global positioning system for precise timing.

The beneficial effects of the present invention include:

The invention provides a method for reducing the false alarm rate of pipeline leakage monitoring. It judges whether the peak ratio of the abnormal signal received by the acoustic wave monitor at the first and last stations is within the preset range by establishing a sound wave propagation attenuation model in the pipeline. Whether the received abnormal signal is a homologous signal. And prevent the system from sending a false alarm when it is judged that it is not a homologous signal. The false alarm rate of existing pipeline leakage monitoring can be effectively reduced, and the reliability of leakage monitoring can be improved, and the method has good robustness.

Description of drawings

Fig. 1 is a flow chart of a specific embodiment of a method for reducing the false alarm rate of pipeline leakage monitoring in the present invention;

Fig. 2 is a flow chart of another specific embodiment of the method for reducing the false alarm rate of pipeline leakage monitoring in the present invention;

Fig. 3 is a schematic diagram of the signal denoising and de-average processing results of the first and last stations of sample 1 of a specific example of a method for reducing the false alarm rate of pipeline leakage monitoring in the present invention;

Fig. 4 is a schematic diagram of the signal denoising and de-average processing results of sample 2 at the first and last stations of a specific example of a method for reducing the false alarm rate of pipeline leakage monitoring in the present invention;

Fig. 5 is a schematic diagram of the inverse Fourier transform result of sample 1 first and last station signals of a specific example of the method for reducing pipeline leakage monitoring false alarm rate after spectrum interception in the present invention;

Fig. 6 is a schematic diagram of the inverse Fourier transform results of the first and last station signals of sample 2 after spectrum interception in a specific example of a method for reducing the false alarm rate of pipeline leakage monitoring according to the present invention.

detailed description

In order to make the purpose, technical solution and advantages of the present invention clearer, the specific implementation of the method for reducing the false alarm rate of pipeline leakage monitoring according to the present invention will be described below with reference to the accompanying drawings. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

The method for reducing the false alarm rate of pipeline leakage monitoring according to an embodiment of the present invention, as shown in Figure 1, comprises the following steps:

S100, respectively, real-time and continuous monitoring of the acoustic wave signal of the first station and the acoustic wave signal of the last station inside the pipeline through the first acoustic wave monitor installed at the first station of the pipeline and the second acoustic wave monitor at the last station of the pipeline.

S200. When an abnormal signal is detected, determine the positioning position of the abnormal signal by calculating the signal correlation coefficient, and respectively calculate the abnormal signal peak value in the sound wave signal of the first station and the abnormal signal peak value in the sound wave signal of the last station.

What needs to be explained here is that the acoustic wave monitor (including the first acoustic wave monitor and the second acoustic wave monitor) continuously monitors the acoustic wave signal. Generally, when there is no leakage in the pipeline, the received signal is relatively gentle without obvious fluctuations. . However, if there is a leak in the monitored pipeline or a strong interference signal (such as pump adjustment, valve adjustment, etc.), the signal received by the acoustic wave monitor will fluctuate significantly, then it is said that the acoustic wave monitor has detected an abnormal signal. . Moreover, the judgment of the abnormal signal can be implemented in combination with the existing technology, and will not be described in detail here.

S300. Determine whether the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is within a preset range. The abnormal signal of the last station is a non-homologous signal, and no leakage alarm is issued.

For any certain transmission pipeline, when homologous abnormal signals appear, the peak ratio between the peak values of the abnormal signals monitored by the acoustic wave monitors set at the first station and at the last station is within a certain range, Therefore, it can be judged whether the received abnormal signal is a homologous signal by judging whether the peak ratio of the abnormal signal received by the acoustic wave monitor at the first and last station is within a preset range. And prevent the system from sending a false alarm when it is judged that it is not a homologous signal. The false alarm rate of existing pipeline leakage monitoring can be effectively reduced, and the reliability of leakage monitoring can be improved.

It should also be noted that the first station here refers to the initial position where the conveyed material flows inside the conveying pipeline and a monitoring point is set here. The end station is the terminal position of the flow of the conveyed material inside the conveying pipeline and a monitoring point is set here. The distance between the first station and the last station is basically the length of the conveying pipeline. Of course, with reference to the method of the present invention, two points close to the first and last stations and with a fixed distance apart can be selected for monitoring on the pipeline.

In one of the embodiments, in step S100, the acoustic wave signal of the first station and the sound wave of the last station inside the pipeline are respectively monitored in real time and continuously by the first acoustic wave monitor installed at the first station of the pipeline and the second acoustic wave monitor installed at the last station of the pipeline. signal, including the following steps:

S110: Sampling the two signals of the first station sound wave signal and the last station sound wave signal output by the first sound wave monitor and the second sound wave monitor respectively with a sampling period T, and performing the collected signals at every minute The two signals are time-stamped respectively.

Among them, the monitoring time of the acoustic wave monitors at the first station and the last station can be accurately timed by GPS (Global Positioning System, Global Positioning System), so as to ensure the synchronization of the acquisition of the acoustic wave signals of the first station and the last station. Improve the accuracy of monitoring.

S120, assuming that the length of the collected signal per minute is N/2 points, respectively extracting the acoustic wave signal of the first station and the acoustic wave signal of the last station for two consecutive minutes, wherein the signal of the previous minute is historical data, and the signal of the next minute is real-time data , the historical data and the real-time data constitute a complete frame of the first-station acoustic signal and the last-station acoustic signal to be processed. Wherein, T is a positive number greater than 0, and N is a positive integer.

Specifically, the preset range of the ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is realized by constructing an acoustic wave propagation attenuation model in the pipeline in the present invention. It diagnoses and locates the abnormal signals of the collected signals, and screens and obtains the abnormal signals forwarded from the first station to the last station and the abnormal signals transmitted backwards from the last station to the first station. Utilize the abnormal signals of the forward and backward transmission mentioned above and select some normal signals, perform discrete Fourier transform after denoising and de-averaging to obtain the spectrum of each signal, compare the spectrum of the abnormal signal with the normal signal, and determine that it is different from the normal signal The abnormal signal band range of the signal. The frequency spectrum of the abnormal signal is filtered in the frequency domain according to the determined frequency band range, and the reconstructed signal in the time domain after the frequency domain filtering is obtained by inverse Fourier transform. The reconstructed signal is divided into positive and negative intervals, the peak value is found in the interval where the prominent signal is located, and the ratio of the peak value of the first and last station is calculated. The attenuation coefficient is calculated according to the sound wave propagation attenuation formula, and the sound wave propagation attenuation model in the corresponding pipeline is established.

Specifically, S010, the step of determining the preset range of the ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station includes the following steps:

S011. Obtain multiple groups of first-station sound wave signals and last-station sound wave signals, and perform filtering, denoising, and de-averaging processing to obtain denoised signals.

S012. Obtain the peak position of the correlation coefficient between the abnormal signal at the first station and the abnormal signal at the last station through correlation calculation on the denoised signal, which is denoted as CorrPos. According to the formula:

Calculate the location where the abnormal signal occurs, where L is the total length of the pipeline, Vdn and Vup are the propagation speeds of the acoustic signal from the first station to the last station and from the last station to the first station, respectively. Obtain the abnormal signal of the first station whose positioning position is the first station and the abnormal signal of the last station of the same source, as well as the abnormal signal of the first station and the abnormal signal of the last station of the same source whose positioning position is the last station, and obtain a certain number (the first preset number) The normal signal of the first station and the normal signal of the last station.

It should be noted here that the preset time is determined according to actual needs, and the present invention requires multiple sets of first station abnormal signals and end station abnormal signals of the same source during the process of determining the preset range. Therefore, the preset time here can be determined according to the actual desired precision combined with the complexity of the operation. Similarly, the normal signal is mainly used as a reference, and its specific number can also be selected according to the actual situation.

S013. Perform discrete Fourier transform on the denoised signal to obtain a frequency spectrum of the denoised signal. After discrete Fourier transform, the spectrum X(k) of the signal is obtained:

X(k)=DFT[x(n)]0≤k≤N-1 (2)

In the formula, x(n) is the time-domain signal, and N is the number of discrete Fourier transform points, which is also the total number of signal points.

S014. Compare the frequency spectrum of the normal signal and the abnormal signal, and intercept the frequency spectrum of the abnormal signal. Specifically: draw the frequency spectrum diagram of abnormal signals and normal signals, observe and count the frequency band ranges of each abnormal signal and normal signal, and compare the start and The end sequence numbers are recorded as fst and fend respectively, and the spectrum of this frequency band is reserved for the rest to be cleared. Therefore, a narrower frequency band corresponding to an accurate abnormal signal can be obtained. where the adjusted spectrum is:

The normal signal mentioned here includes the normal signal of the first station and the normal signal of the last station, and the abnormal signal includes the abnormal signal of the first station and the abnormal signal of the last station.

S015. Perform an inverse Fourier transform on the frequency spectrum of the intercepted abnormal signal to obtain a time-domain reconstructed signal of the abnormal signal.

S016. Determine the peak value ratio of the peak value of the abnormal signal at the first station to the peak value of the abnormal signal at the last station according to the amplitude of the reconstructed signal in the time domain.

S017, according to the propagation and attenuation formula of sound waves along the pipeline:

Among them, Peak ₀ is the initial amplitude of the acoustic signal at the leakage point, Peak is the signal amplitude after propagation attenuation, l is the distance between the signal occurrence location and the first station, L is the total length of the pipeline, and α is the acoustic wave propagation attenuation coefficient. It is obtained that when the acoustic signal generated by the first station propagates downstream to the last station, or when the acoustic signal generated by the last station propagates countercurrently to the first station, they respectively satisfy:

In the formula, the propagation distance of the acoustic signal is L, Peak _up is the amplitude of the abnormal signal collected by the first station, Peak _dn is the amplitude of the abnormal signal of the same source collected by the last station, α _s and α _n are the downstream and upstream channels respectively The propagation attenuation coefficient when propagating.

S018, further calculation can obtain the attenuation coefficient α _s and the attenuation coefficient α _n of the downstream propagation of the acoustic wave signal as follows:

S019. Obtain specific values of α _s and α _n according to the peak ratio of the abnormal signal at the first station determined in step S016 to the abnormal signal at the end station of the same source.

S020, combining the peak ratios of the second preset number of pairs of abnormal signals at the first station and the abnormal signals at the end station of the same source to determine the ranges of the downstream propagation attenuation coefficient α _s and the upstream propagation attenuation coefficient α _n of the acoustic wave signal as follows:

As an implementable manner, the second preset number may be 10 or 12. Of course, more numbers can also be selected according to the requirements of actual precision.

After determining the attenuation coefficients of forward flow and reverse flow, proceed to the next step, S021, for the leakage acoustic wave signal that occurs at a distance l from the first station, the peak ratio r of the abnormal signal at the first station and the abnormal signal at the end station of the same source satisfies The following relationship:

The sound wave propagation attenuation coefficients are divided into four groups [α _smin ,α _nmin ],[α _smin ,α _nmax ],[α _smax ,α _nmin ],[α _smax ,α _nmax ], and four r values are obtained, and the maximum It is recorded as r _max and the minimum value is recorded as r _min , then the value range of the peak value ratio of the abnormal signal at the first station and the abnormal signal at the last station is:

Finally, it is determined that the ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is between r _min and r _max , and the above formula is also the acoustic signal propagation attenuation model in the pipeline constructed.

As can be seen above, when determining whether the ratio of the abnormal signal peak value determined by the first station to the abnormal signal peak value determined by the last station is within a preset range, it is also necessary to determine Specific values for preset ranges. Correspondingly, in step S300, it is judged whether the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station is within the preset range, if so, the abnormal signal at the first and last station is a homologous signal, and a leak alarm is issued; if No, the abnormal signal of the first and last station is a non-homologous signal, and no leakage alarm is issued, including the following steps:

S310. Calculate the peak ratio between the peak value of the abnormal signal at the first station and the peak value of the abnormal signal at the last station.

S320. Calculate the preset range of the peak ratio according to formula (10) and values of 1, α _s and α _n .

S330, judging whether the peak ratio is within the calculated preset range; if yes, the abnormal signal of the first and last stations is a signal of the same source, and a leak alarm is issued; if not, the abnormal signal of the first and last station is a non-homologous signal signal, no leak alarm.

Specifically, in step S310, calculate the peak ratio between the peak value of the abnormal signal peak value at the first station and the peak value value of the abnormal signal signal at the last station; The peak-to-peak ratio of the peak, the method is the same, including the following steps:

S311, divide the positive and negative intervals according to the polarity of the amplitude, and obtain the peak value of each interval as Peak[i], take the maximum value in the positive interval, and take the minimum value in the negative interval, wherein, 1≤i≤N _Peak , N _Peak is the total number of intervals in a complete frame of the first-station acoustic signal or the last-station acoustic signal to be processed;

S312. Calculate the mean value of the positive peak value and the negative peak value, which are recorded as meanVP and meanVN respectively, and calculate the normalized peak prominence index L _P for the peak value greater than meanVP in the positive signal peak value according to the following formula. When the positive peak value is less than meanVP, let LP is 0,

Calculate the normalized peak prominence index L _N of the negative peak with an amplitude greater than meanVN according to the following formula. When the negative peak is greater than meanVN, set LN to 0,

Clearing the signals of the signal intervals in which the normalized peak prominence index is lower than a preset threshold value to 0, determining the peak interval in which the normalized peak prominence index is higher than a preset threshold value, and determining the time when the signal occurs LeakPos.

What needs to be explained here is that when the peak prominence index is closer to 1 or -1, the signal is more prominent; when the prominence index is close to 0, it means that the signal is not prominent. According to the peak prominence degree index and the preset threshold value, the interval of the peak value of the protruding signal is determined, and the time when the protruding signal occurs is determined. The signal difference in each interval is calculated according to the formula d[i]=x[i]-x[i+w]. The step size of the difference is w, and x is the time-domain signal after denoising. Find the maximum value position in the signal difference d of the outstanding signal and record it as the occurrence time of the current interval signal LeakPos, record the time when the abnormal signal at the first station occurs as LeakPos _up (i), and record the time when the abnormal signal at the last station occurs as LeakPos _dn (j), i and j are positive integers, and both are less than or equal to the total number of intervals of the signal they belong to. All the prominence indicators in the signal of the first and last stations are cleared to 0 in the signal intervals that are lower than the preset threshold.

S313, when there are a plurality of abnormal signals, calculate the difference of the LeakPos of one abnormal signal of the first station and each abnormal signal of the last station in turn:

DT＝LeakPos _up (i)-LeakPos _dn (j) (14)

When the difference DT is the closest to the CorrPos in the aforementioned S012, it is determined that this group of signals is the abnormal signal of the first station and the corresponding abnormal signal of the last station;

S314. Obtain the peak value Peak(Range _up ) of the abnormal signal at the first station and the peak value Peak(Range _dn ) of the abnormal signal at the last station, and further obtain the peak value ratio of the abnormal signal at the first station to the peak value of the abnormal signal at the last station.

What needs to be explained here is that the homology of the abnormal signal is actually monitored after the preset range of the ratio of the peak value of the abnormal signal at the first station to the peak value of the abnormal signal at the last station is determined. During the monitoring process, as shown in Figure 2, after monitoring the abnormal signal, steps S011-S015 are also performed to process the signal and calculate the location of the abnormal signal as the distance l from the first station and the peak position of the correlation coefficient CorrPos. And continue to execute steps S311-S314. As shown in Figure 2, after obtaining the time-domain reconstructed signal, continue to divide the positive and negative intervals according to the polarity of the signal amplitude (S311), find out the time when the abnormal signal occurs according to the positioning result combined with the difference calculation, and obtain the peak value of the abnormal signal. Determine the interval of the abnormal signal according to the peak prominence degree (S312-S313), then perform step S314 to calculate the peak value ratio of the abnormal signal at the first station to the peak value of the abnormal signal at the last station, and finally judge whether the pipeline leaks according to the preset range, If so, a leak alarm will be issued, if not, it will be judged as a non-homologous signal, and no alarm will be issued.

A specific example is given below to illustrate the application of the method of the present invention.

Two abnormal signal samples are selected for identification, sample 1 is an interference signal, sample 2 is a leakage signal, and both samples are a complete signal of one frame. The algorithm flow is shown in Figure 2, and the method of the present invention can be implemented by programming in any language and run on a computer.

Step 1. By making statistics and comparing the spectrum diagrams of the abnormal signals and normal signals of a certain amount of pipelines to be tested, the abnormal signal frequency range is the 5th to 45th spectral lines and the 5967th to 5997th spectral lines in the spectrum ( Symmetry of the Fourier transform spectrum). In the attenuation model established according to historical data, α _smax =-0.2785, α _smin =-0.4884, α _nmax =1.9640, α _nmin =1.5746. Through the abnormal signal diagnosis method, it is diagnosed that there are abnormal signals in the two frames of signals, the signal length is N=6000 points, the sampling rate is 50Hz, and the total length of the pipeline is L=12.409km.

Step 2. Use moving average filtering to denoise, and the scale is 100; after moving average filtering and de-averaging, a bipolar signal of the first and last stations is obtained. Figure 3 shows the signal of the first and last station of sample 1 after processing, and Figure 4 shows the signal of the first and last station of sample 2. The upper part of the two figures is the signal diagram of the first station, and the lower part is the signal diagram of the last station.

Step 3. Calculate the peak value of the correlation coefficient between sample 1 and sample 2, and calculate the location of the abnormal signal of sample 1 as l ₁ =8.6161km, and the location of the abnormal signal of sample 2 as l ₂ =0.6006km

Step 4. In the frequency spectrum of the first and last station signals, all the spectral lines except the frequency band of the abnormal signal mentioned in step 1 are cleared to 0.

Step 5. Inverse Fourier transform the above frequency domain filtered spectrum to obtain the time domain reconstruction waveform, the time domain reconstruction waveform of sample 1 is shown in Figure 5, and the time domain reconstruction waveform of sample 2 is shown in Figure 6 Show.

Step 6. According to the polarity of the signal amplitude, the signals of the first and last stations are divided into several positive and negative intervals. In this example, sample 1 has 46 first-station signal intervals and 49 last-station signal intervals; sample 2 has 37 first-station signal intervals and 55 last-station signal intervals.

Step 7. The abnormal signal of the first station of sample 1 is located in the 37th interval, and the abnormal signal of the last station is located in the 34th interval. The signal in the interval is calculated as a difference with a delay scale of 50. The position of the extreme point after the signal difference of the first station is x ₁ (4248), the last station is y ₁ (3908), the abnormal signal peak value of the first station Peak _{up1 =227.3970, the abnormal signal peak value of the last station Peak dn1 =} 581.4869, and the amplitude ratio is r ₁ =Peak _up1 /Peak _dn1 =0.3911.

The abnormal signal of the first station of sample 2 is located in the 9th interval, and the abnormal signal of the last station is located in the 17th interval. The signal in the interval is calculated as a difference with a delay scale of 50, and the position of the extreme point after the difference of the abnormal signal of the first station is x ₂ ( 957), the last station is y ₂ (1775), the abnormal signal peak value of the first station Peak up2 =181.4235, the abnormal signal peak value of the last station Peak _dn2 = _294.1498 , and the amplitude ratio is r ₂ =Peak up2 /Peak _dn2 = _0.6168 .

Step 8. Substituting the positioning distance between sample 1 and sample 2 into the model, the value range of the amplitude ratio of sample 1 is [0.2203,0.3087], and the value range of the amplitude ratio of sample 2 is [0.5713,0.7109]. If the amplitude ratio r ₂ of sample 2 is within the value interval, it is judged as a homologous leakage signal; if the amplitude ratio r ₁ of sample 1 is outside the value interval, it is judged as a non-homologous interference signal.

The method for reducing the false alarm rate of pipeline leakage monitoring in the present invention is based on the sound wave propagation attenuation model, and is modeled according to the actual signal of the pipeline, thereby greatly reducing the influence of the sensitivity of the sensor and the magnification of the transmitter. Since the time-frequency characteristics of the interference signal are close to those of the leakage signal, it is difficult to distinguish them by the previous pattern recognition method, which may easily cause false positives. However, the method can effectively and quickly realize the identification of pipeline leakage interference signals, improve the accuracy of pipeline leakage monitoring and alarm, and has better robustness.

The above-mentioned embodiments only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but should not be construed as limiting the patent scope of the present invention. It should be pointed out that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention should be based on the appended claims.

Claims (6)

1. it is a kind of reduce line leakage rate of false alarm method, it is characterised in that comprise the following steps：

It is real-time respectively, continuous by the second sound wave monitor for being arranged on the first sound wave monitoring instrument and pipeline terminal of pipeline initial station Initial station acoustic signals and terminal acoustic signals inside ground monitoring pipeline；

When monitoring that initial station and terminal simultaneously when there is abnormal signal, calculate the abnormal letter in initial station in the acoustic signals of initial station respectively Terminal abnormal signal peak value in number peak value and terminal acoustic signals；

The peakedness ratio between the initial station abnormal signal peak value and terminal abnormal signal peak value is judged whether in preset range, if It is that then the pipeline occurs leakage；If it is not, then there is no leakage in the pipeline；

The step of the preset range for also including determining the ratio between the initial station abnormal signal peak value and terminal abnormal signal peak value Suddenly；

Wherein, the preset range of the ratio between the initial station abnormal signal peak value and the terminal abnormal signal peak value is to pass through Build what Acoustic Wave Propagation attenuation model in the pipeline was realized.

2. it is according to claim 1 reduce line leakage rate of false alarm method, it is characterised in that it is described by arrange In real time, continuously monitored in pipeline in the second sound wave monitor of the first sound wave monitoring instrument and pipeline terminal of pipeline initial station respectively The initial station acoustic signals and terminal acoustic signals in portion, comprise the following steps：

With sampling period T sample respectively the first sound wave monitoring instrument and the second sound wave monitor output the initial station sound Ripple signal and the terminal acoustic signals two paths of signals, stamp time mark respectively at whole moment minute to the two paths of signals for gathering Sign；

The continuous two minutes initial station acoustic signals and the terminal acoustic signals are extracted respectively, wherein previous minute signal is to go through History data, one minute after signal is real time data, and the historical data and the real time data constitute complete pending of a frame Initial station acoustic signals and terminal acoustic signals；

Wherein, T is the positive number more than 0.

3. it is according to claim 2 reduce line leakage rate of false alarm method, it is characterised in that described in the determination The step of preset range of the ratio between initial station abnormal signal peak value and terminal abnormal signal peak value, specifically include following step Suddenly：

Obtain multigroup initial station acoustic signals and terminal acoustic signals and be filtered, denoising and remove average value processing, after obtaining denoising Signal；

The peak value of the coefficient correlation of initial station abnormal signal and terminal abnormal signal is obtained to the signal after denoising by correlation computations Position, is denoted as CorrPos, and according to formula：

$1=\frac{V_{up}(L+{\mathrm{CorrPosV}}_{dn})}{V_{up}+V_{dn}},$

Determine that abnormal signal occurs distance of the position far from initial station, wherein, l is that abnormal signal occurs distance of the position far from initial station, and L is Pipeline total length, V_dnWith V_upAcoustic signals spread speed respectively from initial station to terminal and from terminal to initial station；

After the monitoring of Preset Time, initial station abnormal signal and homologous terminal letter extremely that position location is initial station are obtained Number, and the initial station abnormal signal and homologous terminal abnormal signal that position location is terminal, and obtain the first predetermined number Initial station normal signal and terminal normal signal；

Discrete Fourier transform is carried out to the signal after the denoising, the frequency spectrum of the signal after denoising is obtained；

The frequency spectrum of normal signal and abnormal signal is compared, the frequency spectrum of the abnormal signal is intercepted；

The frequency spectrum of the abnormal signal after to intercepting carries out Fourier inversion, after obtaining the time domain reconstruction of the abnormal signal Signal；

The peak value of initial station abnormal signal and the peak of homologous terminal abnormal signal are determined according to the amplitude of the signal after time domain reconstruction The peakedness ratio of value；

According to sound wave along pipeline propagation attenuation formula：

$Peak={\mathrm{Peak}}_0{e}^{-\∝\frac{1}{L}},$

Wherein, Peak₀There is the initial magnitude of point acoustic signals for abnormal sound wave, Peak is the signal amplitude after propagation attenuation, l There is distance of the position far from initial station for abnormal signal, L is pipeline total length, and α is Acoustic Wave Propagation attenuation coefficient, at first stop the sound of generation Ripple signal downstream propagation is to terminal, and terminal produces acoustic signals adverse current when traveling to initial station, and which meets respectively：

${\mathrm{Peak}}_{dn}={\mathrm{Peak}}_{up}{e}^{-{\mathrm{\α}}_s\frac{L}{L}}$

${\mathrm{Peak}}_{up}={\mathrm{Peak}}_{dn}{e}^{-{\mathrm{\α}}_n\frac{L}{L}}$

Peak_upFor the abnormal signal amplitude that initial station collects, Peak_dnFor the homologous abnormal signal amplitude that terminal is collected, α_s With α_nPropagation attenuation coefficient when respectively following current is propagated with adverse current；

Acoustic signals downstream propagation attenuation coefficient α is can be calculated further_sWith adverse current propagation attenuation factor alpha_nFor：

${\mathrm{\α}}_s=-\ln \left(\frac{{\mathrm{Peak}}_{dn}}{{\mathrm{Peak}}_{up}}\right)$

${\mathrm{\α}}_n=-ln\left(\frac{{\mathrm{Peak}}_{up}}{{\mathrm{Peak}}_{dn}}\right)$

α is obtained according to peakedness ratio of the initial station abnormal signal with the homologous abnormal signal of terminal_sAnd α_nValue；

Determine that sound wave is believed with reference to the peakedness ratio of the paired initial station abnormal signal and homologous terminal abnormal signal of the second predetermined number Number downstream propagation attenuation coefficient α_sWith adverse current propagation attenuation factor alpha_nScope be：

α_{s min}≤α_s≤α_{s max}

α_{n min}≤α_n≤α_{n max}；

For generation is in the leakage acoustic signals away from initial station l distances, its initial station abnormal signal and homologous terminal abnormal signal Peakedness ratio r meets following relation：

$r=\frac{{\mathrm{Peak}}_{up}}{{\mathrm{Peak}}_{dn}}=\frac{e^{-{\mathrm{\α}}_n\frac{1}{L}}}{e^{-{\mathrm{\α}}_s\frac{(L-1)}{L}}};$

Acoustic Wave Propagation attenuation coefficient is constituted into four groups of [α_{s min},α_{n min}],[α_{s min},α_{n max}],[α_{s max},α_{n min}],[α_{s max}, α_{n max}], four r values are obtained, and maximum is designated as into r_max, minimum of a value is designated as r_min, then initial station abnormal signal and homologous terminal The span of the peakedness ratio of abnormal signal is：

4. it is according to claim 3 reduce line leakage rate of false alarm method, it is characterised in that judge the initial station Peakedness ratio between abnormal signal peak value and terminal abnormal signal peak value whether in preset range, if so, then send out by the pipeline Raw leakage；If it is not, then the pipeline does not occur leakage, comprise the following steps：

Calculate the peakedness ratio between the initial station abnormal signal peak value and the terminal abnormal signal peak value；

According to formulaAnd α_sWith α_nValue calculate the preset range of the peakedness ratio；

Judge the peakedness ratio whether in the preset range for being calculated；

If so, then first and last station abnormal signal is homologous signal, sends leakage alarm；

If it is not, then first and last station abnormal signal is non-homogeneous signal, leakage alarm is not sent.

5. it is according to claim 3 reduce line leakage rate of false alarm method, it is characterised in that it is described according to time domain The amplitude of the signal after reconstruct determines the peakedness ratio of the peak value of initial station abnormal signal and the peak value of terminal abnormal signal, including following Step：

Positive and negative interval is divided according to the polarity of the amplitude, and obtains each interval peak value and be designated as Peak [i], positive interval takes Maximum, takes minimum of a value between minus zone, wherein, 1≤i≤N_Peak, N_PeakFor the complete pending initial station acoustic signals of a frame or end The interval sum stood in acoustic signals；

The average of positive peak and negative peak is calculated, meanVP and meanVN is designated as respectively, and will be more than in positive signal peak value The peak value of meanVP calculates normalization peak value projecting degree index L according to equation below_P, and when positive peak is seasonal less than meanVP L_PFor 0,

$L_p=\frac{Peak\[i\]-meanVP}{Peak\[i\]},1\≤i\≤N_{Peak},Peak\[i\]>0;$

Peak value of the amplitude in negative peak more than meanVN is calculated into normalization peak value projecting degree index L according to equation below_N, and When negative peak is more than meanVN season L_NFor 0,

$L_N=\frac{Peak\&lsqb;i\&rsqb;-meanVN}{Peak\&lsqb;i\&rsqb;},1\&le;i\&le;N_{Peak},Peak\&lsqb;i\&rsqb;<0;$

The signal that the normalization peak value projecting degree index is less than the signal spacing of predetermined threshold value is 0 clearly, it is determined that described return One changes peak value be located interval and generation moment LeakPos of the peak value projecting degree index higher than predetermined threshold value；

When there are multiple abnormal signals, one abnormal signal of initial station and the LeakPos's of each abnormal signal of terminal are calculated successively Difference：

DT=LeakPos_up(i)-LeakPos_dn(j)

When difference DT and the CorrPos closest to when, determine that this group of signal is respectively initial station abnormal signal and corresponding terminal Abnormal signal, now interval sequence number be designated as Range respectively_upWith Range_dn；

And then obtain the peak value Peak (Range of initial station abnormal signal_up) and terminal abnormal signal peak value Peak (Range_dn), and The peak value of the initial station abnormal signal and the peakedness ratio of the terminal abnormal signal are obtained further.

6. it is according to claim 2 reduce line leakage rate of false alarm method, it is characterised in that using pipeline initial station The first sound wave monitoring instrument and pipeline terminal the second sound wave monitor monitoring pipeline inside initial station acoustic signals and terminal sound During ripple signal, precision time service is carried out using global positioning system.