CN117871466A - Laser methane detector for monitoring data self-checking - Google Patents
Laser methane detector for monitoring data self-checking Download PDFInfo
- Publication number
- CN117871466A CN117871466A CN202410275303.8A CN202410275303A CN117871466A CN 117871466 A CN117871466 A CN 117871466A CN 202410275303 A CN202410275303 A CN 202410275303A CN 117871466 A CN117871466 A CN 117871466A
- Authority
- CN
- China
- Prior art keywords
- time sequence
- data
- current
- data window
- window
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 title claims abstract description 228
- 238000012544 monitoring process Methods 0.000 title claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 61
- 238000001914 filtration Methods 0.000 claims abstract description 59
- 230000005856 abnormality Effects 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 9
- 238000000034 method Methods 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 6
- 238000007689 inspection Methods 0.000 claims description 6
- 230000002441 reversible effect Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 abstract description 11
- 230000000694 effects Effects 0.000 abstract description 2
- 238000000354 decomposition reaction Methods 0.000 description 13
- 230000000737 periodic effect Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 3
- 238000010521 absorption reaction Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- 238000000862 absorption spectrum Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to the technical field of laser gas detection, in particular to a laser methane detector for monitoring data self-detection, which comprises a laser methane sensor and a data processor, wherein the laser methane sensor detects laser methane concentration data, the data processor intercepts a plurality of time sequence data windows from the laser methane concentration time sequence data, obtains abnormal fluctuation factors according to the current period data window and the data correlation condition in each time sequence data window, further obtains the current noise-containing constraint factor corresponding to the current time sequence data window, combines the obtained data to obtain the confidence factor of the current time sequence data window, further obtains the filtering adjustment coefficient of the current data point, adopts an SG filtering algorithm to carry out data filtering, improves the filtering effect of noise, simultaneously avoids the damage to information in an original data curve as much as possible, improves the accuracy of data filtering, and further enables judgment to be more accurate when judging the methane gas concentration.
Description
Technical Field
The invention relates to the technical field of laser gas detection, in particular to a laser methane detector for monitoring data self-detection.
Background
The methane gas concentration is usually detected by using a laser methane detector, and the requirement on the detection precision of the laser methane detector is higher for some occasions with higher safety requirements. However, due to the limitation of the sensor, noise interference is usually caused in the collected methane concentration data, so in order to improve the detection precision and avoid the condition of missing report and false report, SG filtering denoising is usually performed on the collected methane concentration data.
However, in the actual monitoring process, because of the reasons of fluidity, diffusivity, instability of the laser transmission path and the like of the methane gas in the gas pool of the laser methane detector in addition to noise data, fluctuation exists in the obtained concentration data curve of the actual methane gas, but the fluctuation is allowable.
Since the common SG filtering generally adopts a window with a fixed size, and fits the data in the window, so as to complete filtering and smoothing denoising. However, due to the existence of the normally allowable fluctuation, the existing SG filtering method cannot accurately denoise methane gas detection data, and needs to adopt higher filtering intensity for data in a local range with noise interference and serious noise interference, and adopts smaller filtering intensity for data of a part not interfered by the noise, so that more detailed information of the original data is reserved.
Disclosure of Invention
In view of the above, the invention provides a laser methane detector for self-checking monitoring data in order to solve the technical problem that the data filtering mode in the existing laser methane detector can not accurately denoise methane gas detection data.
The adopted technical scheme is as follows:
the laser methane detector for monitoring the data self-detection comprises a laser methane sensor and a data processor, wherein the laser methane sensor is in signal connection with the data processor, the laser methane sensor is used for detecting laser methane concentration data according to a preset sampling period and transmitting the detected laser methane concentration time sequence data to the data processor, and the data processor performs the following data processing process:
intercepting a plurality of time sequence data windows from the laser methane concentration time sequence data, wherein the time sequence data windows comprise a current time sequence data window and at least one historical time sequence data window, and the current trend data window corresponds to a current data point;
respectively and correspondingly acquiring a current period abnormal fluctuation factor and each time sequence abnormal fluctuation factor according to a current period data window corresponding to the current time sequence data window, the time interval between extreme points in each time sequence data window and the amplitude difference; acquiring a current noise-containing constraint factor corresponding to a current time sequence data window according to the time sequence abnormal fluctuation factor corresponding to the current time sequence data window, the current period abnormal fluctuation factor, the amplitude range and the extreme point number of the current trend data window corresponding to the current time sequence data window;
Obtaining a confidence factor of the current time sequence data window according to the data difference among the time sequence data windows, each time sequence abnormal fluctuation factor and the residual accumulated value corresponding to each time sequence data window;
acquiring a filtering adjustment coefficient of a current data point according to a time sequence abnormal fluctuation factor, a current noise-containing constraint factor and a confidence factor corresponding to a current time sequence data window;
and according to the filtering adjustment coefficient of the current data point, adopting an SG filtering algorithm to carry out data filtering.
Further, a plurality of time sequence data windows are intercepted from the laser methane concentration time sequence data, and the time sequence data windows comprise:
taking the current data point as a starting point, intercepting and dividing the laser methane concentration time sequence data in a reverse time sequence direction, and sequentially obtaining a plurality of time sequence data windows with preset window lengths, wherein the time sequence data window corresponding to the current data point is the current time sequence data window, and the rest time sequence data windows are the history time sequence data windows;
and taking the data points corresponding to the time sequence end points in each time sequence data window as characteristic data points of the corresponding time sequence data window, wherein the characteristic data points of the current time sequence data window are the current data points.
Further, the obtaining the current period abnormal fluctuation factor and each time sequence abnormal fluctuation factor according to the time interval and the amplitude difference between the current period data window corresponding to the current time sequence data window and the extreme point in each time sequence data window, respectively, includes:
for any time sequence data window, all extreme points in the time sequence data window are obtained, the time interval and the amplitude difference between each extreme point and one extreme point adjacent to the extreme point are obtained, the standard deviation of all the time intervals in the time sequence data window is calculated, the time sequence time interval difference characteristic quantity is obtained, and the standard deviation of all the amplitude differences in the time sequence data window is calculated, so that the time sequence amplitude difference characteristic quantity is obtained;
based on the position of the characteristic data point of the time sequence data window in the time sequence data window, acquiring data windows corresponding to other data points in the time sequence data window in the laser methane concentration time sequence data, and obtaining each associated time sequence data window, wherein the length of each associated time sequence data window is equal to the preset window length; acquiring time sequence time interval difference characteristic quantities and time sequence amplitude difference characteristic quantities of each associated time sequence data window;
Obtaining a time sequence abnormal fluctuation factor of the time sequence data window according to the time sequence time interval difference characteristic quantity and the time sequence amplitude difference characteristic quantity of the time sequence data window and each related time sequence data window;
acquiring all extreme points in a current period data window, acquiring time intervals and amplitude differences between each extreme point and one extreme point adjacent to the extreme point, calculating standard deviations of all time intervals in the current period data window to obtain period time interval difference characteristic quantities, and calculating standard deviations of all amplitude differences in the current period data window to obtain period amplitude difference characteristic quantities;
based on the position of the current data point in the current period data window, acquiring data windows corresponding to other data points in the current period data window, and obtaining each associated period data window, wherein the length of each associated period data window is equal to the preset window length; acquiring cycle time interval difference characteristic quantity and cycle amplitude difference characteristic quantity of each associated cycle data window;
and obtaining the current period abnormal fluctuation factor of the current period data window according to the current period data window, the period time interval difference characteristic quantity and the period amplitude difference characteristic quantity of each associated period data window.
Further, the calculation formula of the timing anomaly fluctuation factor of the timing data window is as follows:
wherein z is a timing anomaly fluctuation factor of the timing data window,l is the number of data points in the time-series data window, which is the normalization function, +.>For the time sequence time interval difference characteristic quantity of the time sequence data window,/for the time sequence time interval difference characteristic quantity of>For the time sequence amplitude difference characteristic quantity of the time sequence data window, < >>For the time sequence time interval difference characteristic quantity of the first other data point in the time sequence data window,/for the first other data point in the time sequence data window>For the first other data point in the time sequence data window, e isNatural constants;
the calculation formula of the current period abnormality fluctuation factor is as follows:
wherein,for the current period abnormality fluctuation factor, +.>Difference feature of period time interval for the current period data window, +.>For the period amplitude difference feature quantity of the current period data window, +.>For the cycle time interval difference feature of the first other data point in the current cycle data window,/v>The cycle amplitude difference feature is the first other data point in the current cycle data window.
Further, according to the timing sequence abnormal fluctuation factor corresponding to the current timing sequence data window, the current period abnormal fluctuation factor, the amplitude range and the extreme point number of the current trend data window corresponding to the current timing sequence data window, the current noise-containing constraint factor corresponding to the current timing sequence data window is obtained, including:
The calculation formula of the current noise-containing constraint factor corresponding to the current time sequence data window is as follows:
wherein c is the current noise-containing constraint factor corresponding to the current time sequence data window, z is the time sequence abnormal fluctuation factor corresponding to the current time sequence data window,for the current period abnormality fluctuation factor, +.>For the number of extreme points of the current trend data window, < +.>For the amplitude range of the current trend data window, +.>For the normalization function, a is a preset parameter greater than 0.
Further, according to the data difference between the time sequence data windows, each time sequence abnormal fluctuation factor and the residual accumulated value corresponding to each time sequence data window, obtaining the confidence factor of the current time sequence data window, including:
the confidence factor of the current time sequence data window is calculated as follows:
wherein g is the confidence factor of the current time sequence data window, z is the time sequence abnormal fluctuation factor corresponding to the current time sequence data window,for the normalized value of the residual accumulated value corresponding to the current time sequence data window, M is the number of historical time sequence data windows, < >>Representing the current time sequence data window and +.>DTW similarity of historical time sequence data window, < >>DTW similarity representing current time sequence data window and all historical time sequence data windows Mean value of->Indicate->Timing anomaly fluctuation factor of each historical timing data window, +.>Is->And the normalized value of the residual accumulated value corresponding to each historical time sequence data window, and q is a preset parameter larger than 0.
Further, according to the timing sequence abnormal fluctuation factor, the current noise constraint factor and the confidence factor corresponding to the current timing sequence data window, obtaining a filtering adjustment coefficient of the current data point comprises the following steps:
the calculation formula of the filter adjustment coefficient of the current data point is as follows:
wherein k is a filtering adjustment coefficient of the current data point, z is a time sequence abnormal fluctuation factor corresponding to the current time sequence data window, c is a current noise-containing constraint factor, g is a confidence factor,is a normalization function.
Further, according to the filtering adjustment coefficient of the current data point, performing data filtering by adopting an SG filtering algorithm, including:
acquiring a filtering adjustment coefficient of a preset number of data points according to the time sequence;
obtaining the square sum of residual correction values of the fitting data points and the original data points according to a polynomial constructed in the SG filtering algorithm to obtain a target error model; the residual correction value of the fitting data point and the original data point is equal to the product of the difference value of the fitting data point and the original data point and the filtering adjustment coefficient of the corresponding data point;
And solving the target error model based on the error model minimum principle to obtain a polynomial, and obtaining a fitting value of the target data point based on the obtained polynomial.
Further, the laser methane detector further comprises an alarm and an electromagnetic valve arranged in the gas pipeline, the data processor is in signal connection with the alarm and the electromagnetic valve, the data processor compares laser methane concentration data obtained after data filtering with a preset methane concentration threshold value, if the laser methane concentration data is higher than the preset methane concentration threshold value, an alarm instruction is output to the alarm, and a closing instruction is output to the electromagnetic valve.
Further, the laser methane detector further comprises a wireless communication module, the data processor is in signal connection with the wireless communication module, and the data processor transmits laser methane concentration data obtained after data filtering to a background data management center through the wireless communication module.
The invention has at least the following beneficial effects: firstly, a plurality of time sequence data windows are intercepted from the time sequence data of the laser methane concentration, then, a current period abnormal fluctuation factor is obtained according to the time interval and the amplitude difference between extreme points in the current period data window obtained by time sequence decomposition of the current time sequence data window, each time sequence abnormal fluctuation factor is obtained according to the time interval and the amplitude difference between the extreme points in each time sequence data window, and then, the current noise-containing constraint factor corresponding to the current time sequence data window is analyzed according to the time sequence abnormal fluctuation factor corresponding to the current time sequence data window, the current period abnormal fluctuation factor, the amplitude range and the number of extreme points of the current trend data window corresponding to the current time sequence data window; the method comprises the steps that the actual fluctuation of gas possibly exists in different time sequence data windows and noise possibly exists in the time sequence data windows, the decomposition quality of time sequence decomposition is restrained by the existence of the noise, the periodic fluctuation in periodic data is normally characterized by the fluctuation of the gas, but the noise is overlapped to cause interference to the situation, so that the confidence level of the current noise-containing constraint factor is needed to be analyzed, the confidence factor of the current time sequence data window is obtained according to the data difference among time sequence data windows, each time sequence abnormal fluctuation factor and the residual accumulated value corresponding to each time sequence data window, finally, the filtering adjustment coefficient of the current data point is obtained according to each obtained data factor, the data filtering is carried out by adopting an SG filtering algorithm according to the filtering adjustment coefficient, the filtering effect of the noise is improved, the damage to information in a raw data curve is avoided as much as possible, the accuracy of the data filtering is improved, and the judgment can be more accurate when the methane gas concentration is judged.
Drawings
FIG. 1 is a schematic diagram of the hardware components of a laser methane detector for monitoring data self-inspection provided by the invention;
FIG. 2 is a flow chart of data processing performed by a data processor in a laser methane detector for monitoring data self-checking, provided by the invention;
FIG. 3 is a schematic diagram of acquisition of a time series data window;
FIG. 4 is a schematic diagram of associated time series data window acquisition for a current time series data window.
Detailed Description
The embodiment provides a laser methane detector for monitoring data self-checking, as shown in fig. 1, which comprises a laser methane sensor and a data processor, wherein the laser methane sensor can be conventional laser methane sensor equipment and is used for detecting methane gas concentration in an application scene through a laser detection principle. The laser detection principle is briefly described as follows: the laser methane sensor is based on a tunable semiconductor laser absorption spectrum technology, the laser transmitter outputs light power, sawtooth current scanning occurs when the light power passes through a gas tank to a reflecting mirror, the reflecting mirror receives light intensity after passing through the gas tank again when returning the light power to the laser receiver, absorption occurs once in each period of the sawtooth wave, the methane gas concentration can be detected according to the size of an absorption peak, the laser receiver is connected with a data processor through signals, and the laser receiver is used for processing to obtain a methane gas concentration value. The data sensor may be a conventional data processing chip such as a single chip microcomputer, a CPU (Central Processing Unit ) or the like. The laser methane sensor is in signal connection with the data processor, can be connected through a signal transmission line, and can also be welded on a PCB circuit board to realize signal transmission. The laser methane sensor is used for detecting laser methane concentration data according to a preset sampling period, forming laser methane concentration time sequence data according to the laser methane concentration data detected by the preset sampling period, and transmitting the detected laser methane concentration time sequence data to the data processor.
As a specific implementation mode, the laser methane detector further comprises an alarm and an electromagnetic valve, wherein the alarm can be a buzzer or an audible and visual alarm, the electromagnetic valve is arranged in a gas pipeline, and the electromagnetic valve has two states of on and off. The data processor is in signal connection with the alarm and the electromagnetic valve. Furthermore, the laser methane detector also includes a wireless communication module, such as: and the 5G module is in signal connection with the data processor and the wireless communication module, so that remote wireless transmission of data is realized.
As shown in fig. 2, the data processor performs the following data processing procedure:
step 1: intercepting a plurality of time sequence data windows from laser methane concentration time sequence data, wherein the time sequence data windows comprise a current time sequence data window and at least one historical time sequence data window, and the current trend data window corresponds to a current data point:
in this embodiment, the preset sampling period of the laser methane sensor is set according to actual needs, for example, 0.1 second, and the sampling duration of the laser methane concentration time sequence data is also set according to actual conditions, for example, the direction from the current moment to the earlier moment, and the methane gas concentration data within 1 hour.
A plurality of time sequence data windows are intercepted from the laser methane concentration time sequence data, wherein the lengths of the obtained time sequence data windows are the same, the lengths of the time sequence data windows are preset window lengths, and the preset window lengths are set according to actual needs, for example, 20, namely, the preset window lengths comprise 20 pieces of data. It should be understood that the number of the obtained time sequence data windows is set according to actual needs, and all the obtained time sequence data windows can be part of the laser methane concentration time sequence data, or the laser methane concentration time sequence data is divided into a plurality of parts, each part is one time sequence data window, that is, all the obtained time sequence data windows are all the laser methane concentration time sequence data. Because the final analysis is the relevant characteristic data of the current data point, namely the data point at the current moment, the obtained time sequence data window is divided into a current time sequence data window and at least one historical time sequence data window, and the current trend data window corresponds to the current data point. In the laser methane concentration time sequence data, the current data point is taken as a starting point, the laser methane concentration time sequence data is intercepted and divided in the reverse time sequence direction, namely the direction from the current moment to the earlier moment, time sequence data windows are sequentially acquired, the length of each time sequence data window is the preset window length, the time sequence data window obtained by first interception is the current time sequence data window and corresponds to the current data point, and each time sequence data window obtained by second interception is the history time sequence data window.
Each time sequence data window comprises a plurality of data points, then one data point is determined from each time sequence data window and used as a characteristic data point corresponding to each time sequence data window, and the characteristic information reflected by the time sequence data window characterizes the characteristic of the corresponding characteristic data point. As a specific embodiment, the rightmost data point in the time sequence data window is taken as the characteristic data point of the corresponding time sequence data window, and the data point at the latest moment (namely the instant end point) in the time sequence data window is taken as the characteristic data point of the corresponding time sequence data window. Then the characteristic data point of the current time series data window is the current data point.
As a specific example, as shown in fig. 3, the current time is time t0, the number of data points in each time series data window is 5, and the arrow in fig. 3 indicates time. In the laser methane concentration time sequence data, taking the current data point corresponding to the time t0 as a starting point, and dividing the first 5 data into a current time sequence data window (time sequence data window S1) in the direction of reverse time sequence; the time corresponding to the sixth data point is the time t5, and 5 data are divided into a first historical time sequence data window (time sequence data window S2) from the data point corresponding to the time t 5; the time corresponding to the 11 th data point is the time t10, and 5 data are divided into a second historical time sequence data window (time sequence data window S3) from the data point corresponding to the time t 10; the time corresponding to the 16 th data point is time t15, and 5 data are divided into a third historical time series data window (time series data window S4) from the data point corresponding to the time t 15. The characteristic data points corresponding to the current time sequence data window are data points corresponding to the moment t0, the characteristic data points corresponding to the first historical time sequence data window are data points corresponding to the moment t5, the characteristic data points corresponding to the second historical time sequence data window are data points corresponding to the moment t10, and the characteristic data points corresponding to the third historical time sequence data window are data points corresponding to the moment t 15.
Step 2: respectively and correspondingly acquiring a current period abnormal fluctuation factor and each time sequence abnormal fluctuation factor according to a current period data window corresponding to the current time sequence data window, the time interval between extreme points in each time sequence data window and the amplitude difference; acquiring a current noise-containing constraint factor corresponding to a current time sequence data window according to the time sequence abnormal fluctuation factor corresponding to the current time sequence data window, the current period abnormal fluctuation factor, the amplitude range and the extreme point number of the current trend data window corresponding to the current time sequence data window:
it should be appreciated that when noise interference is present in one time series data window, the data fluctuations within that time series data window, whether or not there are gas-induced fluctuations within the remaining time series data windows, are relatively more complex and more severe than the other time series data windows, thus requiring a specific quantification of the fluctuations within each data point in the time series data window.
Firstly, according to the time interval and amplitude difference between the current period data window corresponding to the current time sequence data window and extreme points in each time sequence data window, respectively and correspondingly acquiring the current period abnormal fluctuation factor and each time sequence abnormal fluctuation factor, specifically:
For any one time sequence data window, all extreme points in the time sequence data window are acquired, wherein the extreme points comprise minimum value points and maximum value points, and then for any one extreme point, an adjacent extreme point of the extreme points is acquired, in particular an adjacent extreme point on the left side of the extreme point is acquired. The extreme point at the leftmost side does not participate in the operation because the extreme point adjacent to the leftmost side does not exist. Then, since the extreme points adjacent to the maximum point are the minimum points, the extreme points adjacent to the minimum point are the maximum points, and therefore, for any one of the extreme points, the adjacent one of the extreme points is a different type of extreme point. And obtaining the time interval between the extreme point and one extreme point adjacent to the left side of the extreme point, if the first extreme point of the time sequence data window does not have the extreme point on the left side, no time interval exists, and other extreme points have corresponding time intervals, so that a plurality of time intervals corresponding to the time sequence data window are obtained, standard deviations of all the time intervals in the time sequence data window are calculated, and time sequence time interval difference characteristic quantities are obtained. And similarly, acquiring the amplitude difference (namely the methane gas concentration value corresponding to the extreme point) between the extreme point and one extreme point adjacent to the left side of the extreme point, so as to obtain a plurality of amplitude differences corresponding to the time sequence data window, and calculating the standard deviation of all the amplitude differences in the time sequence data window to obtain the time sequence amplitude difference characteristic quantity.
Based on the position of the characteristic data point of the time sequence data window in the time sequence data window, acquiring the data window corresponding to other data points in the time sequence data window in the laser methane concentration time sequence data, and obtaining each associated time sequence data window, wherein the length of each associated time sequence data window is equal to the length of a preset window. Specifically: and taking the preset window length as the window length, and taking the sliding window step length as 1, and acquiring a data window corresponding to other data points in the time sequence data window in the laser methane concentration time sequence data. As a specific embodiment, since the characteristic data point of the time-series data window is the data point on the rightmost side of the time-series data window, as shown in fig. 4, for the time-series data window S1, the other data points are the data point corresponding to the time t1, the data point corresponding to the time t2, the data point corresponding to the time t3 and the data point corresponding to the time t4, respectively, the data window corresponding to the data point corresponding to the time t1 is the data window S1.1, the data window corresponding to the data point corresponding to the time t2 is the data window S1.2, the data window corresponding to the data point corresponding to the time t3 is the data window S1.3, and the data window corresponding to the data point corresponding to the time t4 is the data window S1.4. The time sequence data window S1 is correspondingly associated, other data points are respectively data points corresponding to the time t1, data points corresponding to the time t2 and associated time sequence data windows corresponding to the time t3 are respectively data windows S1.1, S1.2, S1.3 and S1.4, and characteristic data points corresponding to the associated time sequence data windows are data points corresponding to the rightmost side of the associated time sequence data windows.
And acquiring the time sequence time interval difference characteristic quantity and the time sequence amplitude difference characteristic quantity of each associated time sequence data window according to the acquisition process of the time sequence time interval difference characteristic quantity and the time sequence amplitude difference characteristic quantity.
Then, according to the time sequence data window and the time sequence time interval difference characteristic quantity and the time sequence amplitude difference characteristic quantity of each related time sequence data window, a time sequence abnormal fluctuation factor of the time sequence data window is obtained, and a calculation formula is as follows:
wherein z is a timing anomaly fluctuation factor of the timing data window,l is the number of data points in the time-series data window, which is the normalization function, +.>For the time sequence time interval difference characteristic quantity of the time sequence data window,/for the time sequence time interval difference characteristic quantity of>For the time sequence amplitude difference characteristic quantity of the time sequence data window, < >>For the time sequence time interval difference characteristic quantity of the first other data point in the time sequence data window,/for the first other data point in the time sequence data window>And e is a natural constant, which is the characteristic quantity of the time sequence amplitude difference of the first other data point in the time sequence data window.
By adopting the mode, the time sequence abnormal fluctuation factors of each time sequence data window are obtained. Z is expressed specifically as a timing anomaly fluctuation factor of the current timing data window as follows.
Since the fluctuation generated by the noise is usually irregular due to the randomness, the distribution among the corresponding extreme points is relatively more disordered, and therefore, when the difference between the time interval and the amplitude difference obtained respectively is large, the more complex the fluctuation in the time sequence data window is, the more abnormal the fluctuation is, and the noise interference is higher.
Moreover, the actual fluctuation of the gas is generally longer in duration than the fluctuation of the noise, so that the fluctuation difference obtained by calculating the current data point and each other data point in the time sequence data window is calculated, and when the difference is smaller, the characteristic that the fluctuation is relatively long in duration with similar fluctuation characteristics is that the corresponding possibility is biased to the possibility that the actual fluctuation characteristics exist, and therefore, certain weakening is needed based on the information.
In this embodiment, the STL time sequence decomposition algorithm is used to decompose the laser methane concentration time sequence data to obtain trend data, period data and residual data, and the trend data, period data and residual data obtained by performing data decomposition according to the STL time sequence decomposition are not described in detail in the prior art.
Therefore, according to the decomposed laser methane concentration time sequence data, a period data window, a trend data window and a residual value corresponding to each time sequence data window are obtained. The period data window corresponding to the current time sequence data window is the current period data window, and the trend data window corresponding to the current time sequence data window is the current trend data window.
According to the acquisition process of the time sequence abnormal fluctuation factor of the time sequence data window, all extreme points in the current period data window are acquired, each extreme point comprises a minimum value point and a maximum value point, the time interval between each extreme point and one extreme point adjacent to the extreme point and the amplitude difference are acquired, wherein for any extreme point, one extreme point adjacent to the extreme point, particularly one extreme point adjacent to the left side of the extreme point, are acquired. The extreme point at the leftmost side does not participate in the operation because the extreme point adjacent to the leftmost side does not exist. Then, since the extreme points adjacent to the maximum point are the minimum points, the extreme points adjacent to the minimum point are the maximum points, and therefore, for any one of the extreme points, the adjacent one of the extreme points is a different type of extreme point. And obtaining the time interval between the extreme point and one extreme point adjacent to the left side of the extreme point, so that the first extreme point of the current period data window has no time interval when the extreme point on the left side does not exist, and other extreme points have corresponding time intervals, thereby obtaining a plurality of time intervals corresponding to the current period data window, calculating standard deviations of all the time intervals in the current period data window, and obtaining period time interval difference characteristic quantities. And similarly, acquiring the amplitude difference between the extreme point and one extreme point adjacent to the left side of the extreme point, so as to obtain a plurality of amplitude differences corresponding to the current period data window, and calculating the standard deviation of all the amplitude differences in the current period data window to obtain the period amplitude difference characteristic quantity.
Based on the position of the current data point in the current period data window, acquiring period data windows corresponding to other data points in the current period data window to obtain each associated period data window, and adopting a division mode shown in fig. 4, wherein the length of each associated period data window is equal to the length of a preset window; and further, acquiring the cycle time interval difference characteristic quantity and the cycle amplitude difference characteristic quantity of each associated cycle data window by adopting the cycle time interval difference characteristic quantity and cycle amplitude difference characteristic quantity acquisition process of the current cycle data window.
According to the current period data window and the period time interval difference characteristic quantity and the period amplitude difference characteristic quantity of each associated period data window, the current period abnormal fluctuation factor of the current period data window is obtained, and the calculation formula is as follows:
wherein,for the current period abnormality fluctuation factor, +.>Difference feature of period time interval for the current period data window, +.>For the period amplitude difference feature quantity of the current period data window, +.>For the cycle time interval difference feature of the first other data point in the current cycle data window,/v>The cycle amplitude difference feature is the first other data point in the current cycle data window.
The above-mentioned content is based on the fluctuation anomaly in the data and relatively intuitively analyzes, but the portion where noise and actual fluctuation coexist needs to be further distinguished. Because the actual fluctuation of the gas is accompanied by the overall trend change, and the change is reflected in the principle of time sequence decomposition, part of the fluctuation is lost, and is reflected in a trend curve instead of a periodic curve after fitting, but noise is not related to the trend characteristics which actually appear due to the randomness of the noise, so that the correlation cannot be generated with the trend in the time sequence decomposition in general, and part of the fluctuation cannot or only has little change. Based on the information, the abnormal fluctuation factor of the time sequence window needs to be restrained, so that the restraint factor is obtained.
And acquiring the amplitude range and the number of extreme points of the current trend data window corresponding to the current time sequence data window, wherein the amplitude range is the difference value between the maximum data and the minimum data in the current trend data window, and the number of the extreme points is the total number of the maximum points and the minimum points contained in the current trend data window.
Then, according to the time sequence abnormal fluctuation factor corresponding to the current time sequence data window, the current period abnormal fluctuation factor, the amplitude range and the extreme point number of the current trend data window corresponding to the current time sequence data window, the current noise-containing constraint factor corresponding to the current time sequence data window is obtained, and the calculation formula is as follows:
Wherein c is the current noise-containing constraint factor corresponding to the current time sequence data window,for the number of extreme points of the current trend data window (+.>Represents a trend data window), the smaller the number of the extreme points, the more single the trend curve change in the current range is considered. />The combination mode of directly multiplying the amplitude range of the current trend data window and the number of extreme points of the current trend data window is considered to be that when the current trend data window corresponding to the current time sequence data window, namely the current trend curve change is more single, the current trend data window is considered to be more singleThe more obvious the characteristics of the overall change currently existing, the weaker the situation that the trend is influenced due to noise interference exists, but the specific influence cannot be represented only based on the change times of the trend change, so the more the trend change times are, the smaller the amplitude change is, the weaker the interference of the noise is still considered, and further the constraint on the abnormal fluctuation factor obtained above is larger. a is a preset parameter greater than 0, which is used to prevent the denominator from being 0, and the specific value of a is set according to actual needs, in this embodiment, a takes a small positive number, such as 0.01.
Time sequence abnormality fluctuation factor z corresponding to current time sequence data window in current noisy constraint factor c calculation formula and current period abnormality fluctuation factorWhen the difference between the abnormal fluctuation factor corresponding to the current periodic data window and the abnormal fluctuation factor corresponding to the current time sequence data window is larger, at this time, it is considered that the partial fluctuation corresponding to the noise is lost from the periodic curve due to the periodic characteristics of the gas fluctuation under the long time sequence considered in the time sequence decomposition and is decomposed into a trend curve or a residual term, so that the larger the difference is considered here, the stronger the noise interference existing in the current time sequence data window is, the weaker the corresponding constraint is, and the logical sequence is in negative correlation, so that the partial fluctuation corresponding to the noise is divided here. Finally, when the obtained noise-containing constraint factor is higher, the characteristic is obtained from the gas fluctuation periodic characteristics existing under the time sequence data of the methane gas concentration, wherein the interference of noise is relatively weak, more fluctuation caused by actual gas change possibly exists, and therefore, the abnormal fluctuation factor of the current data point obtains higher constraint.
Step 3: obtaining a confidence factor of the current time sequence data window according to the data difference among the time sequence data windows, each time sequence abnormal fluctuation factor and the residual accumulated value corresponding to each time sequence data window:
The current noise-containing constraint factor c obtained in the step 2 is obtained based on the decomposed data curve, but the same can exist gas actual fluctuation and noise inside different time sequence data windows, the existence of the noise can constrain the decomposition quality of time sequence decomposition, the periodic fluctuation in the periodic curve is mainly characterized by the fluctuation of the gas, and the superposition of the noise can cause interference to the situation, so that the confidence degree of the current noise-containing constraint factor c of the current time sequence data window is analyzed.
In this embodiment, the number of historical time sequence data windows is set to be M, and DTW (dynamic time warping) algorithm is adopted to calculate the DTW similarity between the current time sequence data window and each historical time sequence data window, i.e. the time sequence similarity. Obtaining a residual value corresponding to the current time sequence data window, calculating an accumulated value of the residual value corresponding to the current time sequence data window, and normalizing the obtained residual accumulated value. And similarly, acquiring the residual value corresponding to each historical time sequence data window, calculating the accumulated value of the residual value corresponding to each historical time sequence data window, and normalizing the residual accumulated value of each historical time sequence data window.
The confidence factor for the current time series data window is calculated as follows:
wherein,confidence factor for the current time series data window, +.>Normalized value of residual accumulated value corresponding to current time sequence data window,/for>Representing the current time sequence data window and +.>DTW similarity of historical time sequence data window, < >>Mean value of DTW similarity representing current time sequence data window and all history time sequence data window, +.>Indicate->Timing anomaly fluctuation factor of each historical timing data window, +.>Is->Normalized values of residual accumulated values corresponding to the historical time series data windows. q is a preset parameter greater than 0, for preventing the denominator from being 0, and the specific value of q is set according to actual needs, in this embodiment, q takes a small positive number, for example, 0.01.
By calculating the current time sequence data window and the first time sequence data windowDTW similarity of each historical time series data window +.>Mean value of DTW similarity with current time sequence data window and all history time sequence data window +.>When the common difference of the DTW similarity between the current time sequence data window and the plurality of historical time sequence data windows is larger, the fluctuation characteristics inside the current time sequence data window are considered to be larger than the gas fluctuation under the normal condition, and the corresponding situation that the periodic curve is abnormal due to noise is more serious.
The timing anomaly fluctuation factor z of the current time series data window is a case where the fluctuation of the current time series data window is considered to have noise interference as well, and is therefore based on this feature as a weight constraint. Normalized value of residual accumulated value corresponding to current time sequence data windowThe residual items used for representing the current time sequence data window are more, the fluctuation characteristic in the current time sequence data window is also represented to be larger than the fluctuation difference under the laser methane concentration time sequence data, so that more residual items can be obtained to be close to the normal fluctuation characteristic, and the residual items are also used as weight coefficients. The time sequence abnormality fluctuation factor z of the current time sequence data window and the normalized value of the residual accumulated value corresponding to the current time sequence data window>Constraint values of the overall confidence factor obtained as the current time series data window.
Finally, the confidence factor of the current time sequence data windowThe larger the time sequence data window is, the periodic fluctuation characteristic of the trend curve after the decomposition of the current time sequence data window is similar under the time sequence data of the laser methane concentration, the corresponding decomposition quality is relatively higher, namely the condition of being interfered by noise is relatively weaker, so that the confidence of the abnormal fluctuation factor is relatively higher.
Step 4: acquiring a filtering adjustment coefficient of a current data point according to a time sequence abnormal fluctuation factor, a current noise-containing constraint factor and a confidence factor corresponding to a current time sequence data window:
because the characteristic information reflected by the time sequence data window characterizes the characteristic of the corresponding characteristic data point, the time sequence abnormal fluctuation factor of the time sequence data window is the time sequence abnormal fluctuation factor of the corresponding characteristic data point. Then, since the characteristic data point corresponding to the current time series data window is the current data point, the time series abnormality fluctuation factor corresponding to the current time series data window is the time series abnormality fluctuation factor of the current data point.
The larger the abnormal fluctuation factor of the time sequence of the current data point is, the stronger the filtering strength of the current data point is considered to be needed, and the larger the noise-containing constraint factor is, the normal fluctuation exists in the current time sequence data window, so that the filtering strength is required to be constrained, the excessive information loss is avoided, and the noise-containing constraint factor is constrained based on the confidence factor.
The calculation formula of the filter adjustment coefficient of the current data point is as follows:
wherein k is a filtering adjustment coefficient of the current data point, z is a time sequence abnormal fluctuation factor corresponding to the current time sequence data window, c is a current noise-containing constraint factor, g is a confidence factor, Is a normalization function.
When k is smaller, the more the current data point participates in the under-band constraint represented by the error in fitting, and then the influence on the fitting parameter optimization result is weaker when fitting parameter optimization is performed subsequently.
Step 5: and according to the filtering adjustment coefficient of the current data point, adopting an SG filtering algorithm to carry out data filtering:
through the steps, the filter adjustment coefficient of the current data point can be obtained, then, as time goes by, the current data point at the current moment changes every time a sampling period passes, and then, the filter adjustment coefficient of the current data point is obtained every time a sampling period passes, and then, the filter adjustment coefficient of a preset number of data points can be obtained according to time sequence. To facilitate matching with the SG filtering algorithm, the filter adjustment coefficients for 2t+1 data points are continuously obtained. At the same time, or at the 2t+1 data points, the laser methane gas concentration was actually detected. And taking the t-th data point (namely, the central data point) as a target data point for curve fitting, and finally obtaining a fitting value of the target data point, namely, the methane gas concentration value after SG filtering.
Setting the filter adjustment coefficient of 2t +1 data points to Representing the filter adjustment coefficient for the i-th data point. Set 2t +1The laser methane gas concentration actually detected by the data points is array x [ i ]]，i=-t、-t+1、…、0、…、t-1、t。
According to SG filtering algorithm, constructing an n-order polynomial, wherein n is less than or equal to 2t+1, to fit the array x [ i ]:
and obtaining a square sum of residual correction values of the fitting data point and the original data point according to the constructed polynomial, and obtaining a target error model, wherein the residual correction values of the fitting data point and the original data point are equal to the product of the difference value of the fitting data point and the original data point and the filtering adjustment coefficient of the corresponding data point. The target error model E is as follows:
and solving the target error model E based on the error model minimum principle, namely using a least square algorithm, and if the fitting result is the best, the square sum in the target error model E is the smallest, namely, the coefficient bias of the polynomial is calculated to be 0 by the E. The coefficients of the polynomial can be obtained by solving, and finally the polynomial is obtained.
And obtaining a fitting value of the t data point (namely a central data point) according to the fitted polynomial, namely the concentration value of the methane gas after SG filtering. By adopting the process, the fitting value corresponding to each sampling time can be obtained.
In this embodiment, the data processor compares the laser methane concentration data obtained after each data filtering, that is, the fitting value, with a preset methane concentration threshold, if the data processor is higher than the preset methane concentration threshold, which indicates that the methane gas concentration is too high and belongs to an abnormal condition, an alarm instruction is output to the alarm, the alarm is controlled to alarm, and a closing instruction is output to the electromagnetic valve, so that the electromagnetic valve is controlled to close, and safety is ensured. In addition, the laser methane detector can also comprise a fan, the fan is arranged on the ventilation window, if the concentration of methane gas is too high, the data processor outputs a starting instruction to the fan, the fan is controlled to start, the indoor methane gas is discharged outdoors, and the user is prevented from being poisoned by sucking methane or from being in explosion accidents due to inflammability.
And the data processor transmits the laser methane concentration data obtained after the data filtering to a background data management center through the wireless communication module.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the scope of the embodiments of the present application, and are intended to be included within the scope of the present application.
Claims (10)
1. The utility model provides a laser methane detector of monitoring data self-checking, includes laser methane sensor and data processor, laser methane sensor with data processor signal connection, laser methane sensor is used for detecting laser methane concentration data according to predetermineeing sampling period to with the laser methane concentration time sequence data who detects is transmitted to data processor, its characterized in that, data processor carries out following data processing process:
intercepting a plurality of time sequence data windows from the laser methane concentration time sequence data, wherein the time sequence data windows comprise a current time sequence data window and at least one historical time sequence data window, and the current trend data window corresponds to a current data point;
respectively and correspondingly acquiring a current period abnormal fluctuation factor and each time sequence abnormal fluctuation factor according to a current period data window corresponding to the current time sequence data window, the time interval between extreme points in each time sequence data window and the amplitude difference; acquiring a current noise-containing constraint factor corresponding to a current time sequence data window according to the time sequence abnormal fluctuation factor corresponding to the current time sequence data window, the current period abnormal fluctuation factor, the amplitude range and the extreme point number of the current trend data window corresponding to the current time sequence data window;
Obtaining a confidence factor of the current time sequence data window according to the data difference among the time sequence data windows, each time sequence abnormal fluctuation factor and the residual accumulated value corresponding to each time sequence data window;
acquiring a filtering adjustment coefficient of a current data point according to a time sequence abnormal fluctuation factor, a current noise-containing constraint factor and a confidence factor corresponding to a current time sequence data window;
and according to the filtering adjustment coefficient of the current data point, adopting an SG filtering algorithm to carry out data filtering.
2. The laser methane detector of claim 1, wherein capturing a plurality of time series data windows from the laser methane concentration time series data comprises:
taking the current data point as a starting point, intercepting and dividing the laser methane concentration time sequence data in a reverse time sequence direction, and sequentially obtaining a plurality of time sequence data windows with preset window lengths, wherein the time sequence data window corresponding to the current data point is the current time sequence data window, and the rest time sequence data windows are the history time sequence data windows;
and taking the data points corresponding to the time sequence end points in each time sequence data window as characteristic data points of the corresponding time sequence data window, wherein the characteristic data points of the current time sequence data window are the current data points.
3. The laser methane detector for monitoring data self-inspection according to claim 2, wherein the acquiring the current period anomaly fluctuation factor and each timing anomaly fluctuation factor respectively according to the time interval and the amplitude difference between the current period data window corresponding to the current timing data window and the extreme point in each timing data window comprises:
for any time sequence data window, all extreme points in the time sequence data window are obtained, the time interval and the amplitude difference between each extreme point and one extreme point adjacent to the extreme point are obtained, the standard deviation of all the time intervals in the time sequence data window is calculated, the time sequence time interval difference characteristic quantity is obtained, and the standard deviation of all the amplitude differences in the time sequence data window is calculated, so that the time sequence amplitude difference characteristic quantity is obtained;
based on the position of the characteristic data point of the time sequence data window in the time sequence data window, acquiring data windows corresponding to other data points in the time sequence data window in the laser methane concentration time sequence data, and obtaining each associated time sequence data window, wherein the length of each associated time sequence data window is equal to the preset window length; acquiring time sequence time interval difference characteristic quantities and time sequence amplitude difference characteristic quantities of each associated time sequence data window;
Obtaining a time sequence abnormal fluctuation factor of the time sequence data window according to the time sequence time interval difference characteristic quantity and the time sequence amplitude difference characteristic quantity of the time sequence data window and each related time sequence data window;
acquiring all extreme points in a current period data window, acquiring time intervals and amplitude differences between each extreme point and one extreme point adjacent to the extreme point, calculating standard deviations of all time intervals in the current period data window to obtain period time interval difference characteristic quantities, and calculating standard deviations of all amplitude differences in the current period data window to obtain period amplitude difference characteristic quantities;
based on the position of the current data point in the current period data window, acquiring data windows corresponding to other data points in the current period data window, and obtaining each associated period data window, wherein the length of each associated period data window is equal to the preset window length; acquiring cycle time interval difference characteristic quantity and cycle amplitude difference characteristic quantity of each associated cycle data window;
and obtaining the current period abnormal fluctuation factor of the current period data window according to the current period data window, the period time interval difference characteristic quantity and the period amplitude difference characteristic quantity of each associated period data window.
4. A laser methane detector according to claim 3, wherein the calculation formula of the timing anomaly fluctuation factor of the timing data window is as follows:
wherein z is a timing anomaly fluctuation factor of the timing data window,l is the number of data points in the time-series data window, which is the normalization function, +.>For the time sequence time interval difference characteristic quantity of the time sequence data window,/for the time sequence time interval difference characteristic quantity of>For the time sequence amplitude difference characteristic quantity of the time sequence data window, < >>For the time sequence time interval difference characteristic quantity of the first other data point in the time sequence data window,/for the first other data point in the time sequence data window>E is a natural constant, which is the time sequence amplitude difference characteristic quantity of the first other data point in the time sequence data window;
the calculation formula of the current period abnormality fluctuation factor is as follows:
wherein,is different from the current periodOften fluctuating factors (F)>Difference feature of period time interval for the current period data window, +.>For the period amplitude difference feature quantity of the current period data window, +.>For the cycle time interval difference feature of the first other data point in the current cycle data window,/v>The cycle amplitude difference feature is the first other data point in the current cycle data window.
5. The laser methane detector for monitoring data self-inspection according to claim 1, wherein obtaining the current noise-containing constraint factor corresponding to the current time sequence data window according to the time sequence abnormality fluctuation factor corresponding to the current time sequence data window, the current period abnormality fluctuation factor, and the amplitude range and the extreme point number of the current trend data window corresponding to the current time sequence data window comprises:
the calculation formula of the current noise-containing constraint factor corresponding to the current time sequence data window is as follows:
wherein c is the current noise-containing constraint factor corresponding to the current time sequence data window, z is the time sequence abnormal fluctuation factor corresponding to the current time sequence data window,for the current period abnormality fluctuation factor, +.>For the number of extreme points of the current trend data window,for the amplitude range of the current trend data window, +.>For the normalization function, a is a preset parameter greater than 0.
6. The laser methane detector for monitoring data self-inspection according to claim 1, wherein obtaining the confidence factor of the current time sequence data window according to the data difference between the time sequence data windows, each time sequence abnormality fluctuation factor and the residual accumulated value corresponding to each time sequence data window comprises:
The confidence factor of the current time sequence data window is calculated as follows:
wherein g is the confidence factor of the current time sequence data window, z is the time sequence abnormal fluctuation factor corresponding to the current time sequence data window,for the normalized value of the residual accumulated value corresponding to the current time sequence data window, M is the number of historical time sequence data windows, < >>Representing the current time sequence data window and +.>DTW similarity of historical time sequence data window, < >>Representing the current time sequence data window and the current time sequence data windowMean value of DTW similarity of historical time sequence data window, < >>Indicate->Timing anomaly fluctuation factor of each historical timing data window, +.>Is->And the normalized value of the residual accumulated value corresponding to each historical time sequence data window, and q is a preset parameter larger than 0.
7. The laser methane detector for monitoring data self-inspection according to claim 1, wherein obtaining the filter adjustment coefficient of the current data point according to the timing anomaly fluctuation factor, the current noise-containing constraint factor and the confidence factor corresponding to the current timing data window comprises:
the calculation formula of the filter adjustment coefficient of the current data point is as follows:
wherein k is a filtering adjustment coefficient of the current data point, z is a time sequence abnormal fluctuation factor corresponding to the current time sequence data window, c is a current noise-containing constraint factor, g is a confidence factor, Is a normalization function.
8. The laser methane detector for monitoring data self-inspection according to claim 1, wherein the filtering of data with SG filtering algorithm according to the filtering adjustment coefficient of the current data point comprises:
acquiring a filtering adjustment coefficient of a preset number of data points according to the time sequence;
obtaining the square sum of residual correction values of the fitting data points and the original data points according to a polynomial constructed in the SG filtering algorithm to obtain a target error model; the residual correction value of the fitting data point and the original data point is equal to the product of the difference value of the fitting data point and the original data point and the filtering adjustment coefficient of the corresponding data point;
and solving the target error model based on the error model minimum principle to obtain a polynomial, and obtaining a fitting value of the target data point based on the obtained polynomial.
9. The laser methane detector for monitoring data self-checking according to claim 1, further comprising an alarm and an electromagnetic valve arranged in a gas pipeline, wherein the data processor is in signal connection with the alarm and the electromagnetic valve, compares laser methane concentration data obtained by filtering data with a preset methane concentration threshold value, and if the laser methane concentration data is higher than the preset methane concentration threshold value, outputs an alarm instruction to the alarm and outputs a closing instruction to the electromagnetic valve.
10. The laser methane detector for monitoring data self-checking according to claim 1, further comprising a wireless communication module, wherein the data processor is in signal connection with the wireless communication module, and the data processor sends laser methane concentration data obtained by filtering the data to a background data management center through the wireless communication module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410275303.8A CN117871466B (en) | 2024-03-12 | Laser methane detector for monitoring data self-checking |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410275303.8A CN117871466B (en) | 2024-03-12 | Laser methane detector for monitoring data self-checking |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117871466A true CN117871466A (en) | 2024-04-12 |
| CN117871466B CN117871466B (en) | 2024-05-24 |
Family
ID=
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268660A (en) * | 2014-10-13 | 2015-01-07 | 国家电网公司 | Trend recognition method for electric power system predication-like data |
| US20210168019A1 (en) * | 2019-12-02 | 2021-06-03 | Alibaba Group Holding Limited | Time Series Decomposition |
| WO2021164267A1 (en) * | 2020-02-21 | 2021-08-26 | 平安科技（深圳）有限公司 | Anomaly detection method and apparatus, and terminal device and storage medium |
| WO2021179572A1 (en) * | 2020-03-12 | 2021-09-16 | 平安科技（深圳）有限公司 | Operation and maintenance system anomaly index detection model optimization method and apparatus, and storage medium |
| CN115659070A (en) * | 2022-12-28 | 2023-01-31 | 鸿基骏业环保科技有限公司 | Water flow data transmission method based on NB-IOT intelligent water meter |
| CN116108008A (en) * | 2023-04-13 | 2023-05-12 | 山东明远生物科技有限公司 | Decorative material formaldehyde detection data processing method |
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN116818739A (en) * | 2023-08-29 | 2023-09-29 | 天津博霆光电技术有限公司 | Indocyanine green detection method based on optics |
| CN116992393A (en) * | 2023-09-27 | 2023-11-03 | 联通(江苏)产业互联网有限公司 | Safety production monitoring method based on industrial Internet of things |
| CN117057517A (en) * | 2023-10-12 | 2023-11-14 | 国网吉林省电力有限公司长春供电公司 | Efficient processing method and system for electric power data based on digital twin |
| CN117407700A (en) * | 2023-12-14 | 2024-01-16 | 国网山东省电力公司莱芜供电公司 | Method for monitoring working environment in live working process |
| CN117454085A (en) * | 2023-10-27 | 2024-01-26 | 杭州三一谦成科技有限公司 | Vehicle online monitoring method and system |
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268660A (en) * | 2014-10-13 | 2015-01-07 | 国家电网公司 | Trend recognition method for electric power system predication-like data |
| US20210168019A1 (en) * | 2019-12-02 | 2021-06-03 | Alibaba Group Holding Limited | Time Series Decomposition |
| WO2021164267A1 (en) * | 2020-02-21 | 2021-08-26 | 平安科技（深圳）有限公司 | Anomaly detection method and apparatus, and terminal device and storage medium |
| WO2021179572A1 (en) * | 2020-03-12 | 2021-09-16 | 平安科技（深圳）有限公司 | Operation and maintenance system anomaly index detection model optimization method and apparatus, and storage medium |
| CN115659070A (en) * | 2022-12-28 | 2023-01-31 | 鸿基骏业环保科技有限公司 | Water flow data transmission method based on NB-IOT intelligent water meter |
| CN116108008A (en) * | 2023-04-13 | 2023-05-12 | 山东明远生物科技有限公司 | Decorative material formaldehyde detection data processing method |
| CN116702081A (en) * | 2023-08-07 | 2023-09-05 | 西安格蒂电力有限公司 | Intelligent inspection method for power distribution equipment based on artificial intelligence |
| CN116818739A (en) * | 2023-08-29 | 2023-09-29 | 天津博霆光电技术有限公司 | Indocyanine green detection method based on optics |
| CN116992393A (en) * | 2023-09-27 | 2023-11-03 | 联通(江苏)产业互联网有限公司 | Safety production monitoring method based on industrial Internet of things |
| CN117057517A (en) * | 2023-10-12 | 2023-11-14 | 国网吉林省电力有限公司长春供电公司 | Efficient processing method and system for electric power data based on digital twin |
| CN117454085A (en) * | 2023-10-27 | 2024-01-26 | 杭州三一谦成科技有限公司 | Vehicle online monitoring method and system |
| CN117407700A (en) * | 2023-12-14 | 2024-01-16 | 国网山东省电力公司莱芜供电公司 | Method for monitoring working environment in live working process |
Non-Patent Citations (2)
| Title |
|---|
| 汪锦: ""基于时空贝叶斯的植被覆盖对环境变化的响应——以京津风沙源治理工程区为例"", 《中国博士学位论文全文数据库 基础科学辑》, 15 January 2021 (2021-01-15) * |
| 陈昆: ""基于时序分解和多特征融合的短时交通流组合预测方法"", 《万方学位论文库》, 12 October 2023 (2023-10-12) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10704982B2 (en) | Sensor recording analysis apparatus and method | |
| US9146800B2 (en) | Method for detecting anomalies in a time series data with trajectory and stochastic components | |
| CN106297114B (en) | A kind of invader detection method and device | |
| CN116992393B (en) | Safety production monitoring method based on industrial Internet of things | |
| CN116484308B (en) | Data acquisition method based on edge self-adaptive calculation | |
| US20060167640A1 (en) | Apparatus and method for dynamic smoothing | |
| CN113568960B (en) | Real-time analysis method and system for data of Internet of things and computer readable storage medium | |
| CN117871466B (en) | Laser methane detector for monitoring data self-checking | |
| CN116842410B (en) | Intelligent helmet antitheft management method and system based on dynamic perception | |
| CN117871466A (en) | Laser methane detector for monitoring data self-checking | |
| CN117119175B (en) | Underground coal mine air door AI video safety management control system | |
| KR20220132824A (en) | Distribution facility condition monitoring system and method | |
| CN111798638A (en) | Auxiliary system fire information processing method based on information fusion | |
| CN116150592A (en) | Wheelchair intelligent monitoring and early warning method and system based on artificial intelligence | |
| EP3543976A1 (en) | A method for increasing specificity of jamming detection in a home alarm system | |
| CN112583378B (en) | Photoelectric sensing signal reconstruction method and system containing baseline drift and high-frequency noise | |
| CN115041544A (en) | Method and device for detecting abnormality of stamping part and storage medium | |
| US20200160678A1 (en) | Energy efficient seismic intrusion detection | |
| US20070114414A1 (en) | Energy signal detection device containing integrated detecting processor | |
| CN117708748B (en) | Operation monitoring system and method for centrifugal fan | |
| JP3877131B2 (en) | Fire detector and fire detection method | |
| KR20150061954A (en) | Method for monitoring adaptive sound field security of environmental noise | |
| CN109672494A (en) | WIFI signal suppressing method and device in a kind of unmanned plane test platform | |
| CN113534169B (en) | Pedestrian flow calculation method and device based on single-point TOF ranging | |
| CN117439827B (en) | Network flow big data analysis method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |