A data processing method for optical fiber sensor with high-precision monitoring

Abstract

The present invention relates to the field of data processing technology, and in particular to a method for processing data of an optical fiber sensor for high-precision monitoring, the method comprising: using an optical fiber sensor to collect geological data sequences at each moment, extracting each subsequence of the geological data sequence according to the distribution relationship of elements in the geological data sequence, obtaining the trend difference of any subsequence according to the degree of difference between any subsequence and other subsequences, obtaining the temperature mutation degree of temperature data at the corresponding moment and the sudden decrease index of oxygen concentration at the corresponding moment according to the trend difference of the subsequence, combining the temperature mutation degree and the sudden decrease index to obtain the corresponding fluctuation index of the subsequence, and completing the abnormal detection of the data collected by the optical fiber sensor according to the corresponding fluctuation index of the subsequence. The present invention aims to improve the accuracy of abnormal detection of data collected by the optical fiber sensor.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a method for processing optical fiber sensor data for high-precision monitoring. Background Art

[0002] Fiber optic sensor is a sensor that uses optical fiber as a sensing element. It can measure and detect various physical quantities and parameters in the environment through changes in optical signals. Compared with traditional electrical sensors, fiber optic sensors have many advantages, such as high sensitivity, strong anti-interference, no electromagnetic interference, corrosion resistance, long-distance transmission, etc. At present, geological disasters occur frequently. It is particularly important to use high-precision data monitoring methods to collect and analyze data in the geological field, to warn and even avoid the occurrence of geological disasters, and to reduce the losses caused by geological disasters.

[0003] Existing anomaly detection methods for data collected by optical fiber sensors often only analyze the data collected by a single sensor, without combining the correlation characteristics of data collected by different optical fiber sensors, resulting in low accuracy of abnormal data detection and false detection and missed detection. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a high-precision monitoring optical fiber sensor data processing method to solve the existing problems.

[0005] The optical fiber sensor data processing method for high-precision monitoring of the present invention adopts the following technical solution:

[0006] An embodiment of the present invention provides a method for processing optical fiber sensor data for high-precision monitoring, the method comprising the following steps:

[0007] Use optical fiber sensors to collect smoke concentration, temperature, and oxygen concentration at each moment;

[0008] The smoke concentrations at all acquisition moments are combined into a smoke concentration sequence, and each subsequence of the smoke concentration sequence is extracted according to the distribution relationship of the elements in the smoke concentration sequence; the trend difference of any subsequence is obtained according to the difference between any subsequence and other subsequences;

[0009] According to the trend difference of the subsequences and the distribution of the temperature data at the corresponding time, the temperature mutation degree of the temperature data at the corresponding time is obtained; according to the trend difference of the subsequences and the distribution of the oxygen concentration at the corresponding time, the sudden decrease index of the oxygen concentration at the corresponding time is obtained;

[0010] The corresponding fluctuation index of the subsequence is obtained by combining the temperature data at the corresponding moment, the temperature mutation degree of oxygen concentration, the sudden drop index, and the correlation between smoke concentration and temperature and oxygen concentration;

[0011] The anomaly detection of the data collected by the optical fiber sensor is completed according to the corresponding fluctuation index of the subsequence.

[0012] Preferably, extracting each subsequence of the smoke density sequence according to the distribution relationship of elements in the smoke density sequence includes:

[0013] For the smoke concentration sequence, the element with the most occurrences is recorded as MS, all the maximum values ​​in the smoke concentration sequence are obtained, and any maximum value is used as the tail element of the subsequence. The elements in the smoke concentration sequence are obtained in sequence from any maximum value as the elements in the subsequence until the first MS element is traversed to the first element of the subsequence.

[0014] Preferably, the trend difference of any subsequence includes:

[0015] Fit all elements of each subsequence into straight lines, obtain the slope of each straight line, calculate the absolute value of the difference between the slope of the fitted straight line of any subsequence and other subsequences, calculate the absolute value of the difference between the coefficient of variation of any subsequence and other subsequences, calculate the DTW distance between any subsequence and other subsequences, calculate the product of two absolute values ​​of the difference and the DTW distance, calculate the mean of the product between any subsequence and all other subsequences, calculate the information entropy and kurtosis of the stress of the transport tunnel surface at the corresponding time of all elements in any subsequence, use the multiplication result of the information entropy and the kurtosis as the stress mutation factor, and use the product of the stress mutation factor and the mean as the trend difference of any subsequence.

[0016] Preferably, the temperature mutation degree of the temperature data at the corresponding moment includes:

[0017] All temperature data at the corresponding moments of the subsequences with the largest trend difference are combined into a temperature detection sequence, and the maximum score of the temperature detection sequence is determined based on the degree of difference between the elements in the temperature detection sequence;

[0018] The temperature mutation degree of temperature detection sequence c2 is recorded as W c2 , the expression is:

[0019]

[0020] In the formula, G c2 is the maximum score of the temperature detection sequence c2, std c2 is the standard deviation of all elements of the temperature detection sequence, is the standard deviation of all collected temperature data, δ is a parameter adjustment coefficient preset to be greater than 0, c2 i+1 、c2 i are the i+1th and ith elements of the temperature detection sequence respectively, and v is the number of elements contained in the temperature detection sequence.

[0021] Preferably, the method for determining the maximum score of the temperature detection sequence is:

[0022] The absolute value of the difference between the first element and the last element in the temperature detection sequence is calculated, recorded as the first difference absolute value, the absolute value of the difference between the last element in the temperature detection sequence and the maximum value in the temperature detection sequence is calculated, recorded as the second difference absolute value, and the sum of the first difference absolute value and the second difference absolute value is used as the maximum value score of the temperature detection sequence.

[0023] Preferably, the sudden decrease index of oxygen concentration at the corresponding moment includes:

[0024] All oxygen concentration data at the corresponding time of the subsequence with the largest trend difference are combined into an oxygen concentration detection sequence, and the decreasing number of the oxygen concentration detection sequence is determined based on the difference between adjacent elements in the oxygen concentration detection sequence;

[0025] Calculate the mean of all the collected oxygen concentration data, count the number of oxygen concentration data less than the mean, calculate the absolute value of the difference between the mean and the mean of all elements in the oxygen concentration detection sequence, record it as the third absolute value of the difference, calculate the ratio of the number of all the collected oxygen concentration data to the number, and use the product of the ratio, the third absolute value of the difference and the number of decreases as the sudden decrease index of the oxygen concentration detection sequence.

[0026] Preferably, the method for determining the decreasing number of times of the oxygen concentration detection sequence is:

[0027] The difference between each element and the previous element in the oxygen concentration detection sequence is calculated and recorded as a first difference, and the number of all the first differences that are less than or equal to 0 is used as the decreasing number of the oxygen concentration detection sequence.

[0028] Preferably, the corresponding volatility index of the subsequence includes:

[0029] The Pearson correlation coefficients of the subsequence with the largest trend difference and the temperature detection sequence and the oxygen concentration detection sequence are calculated respectively, and the sum of the absolute values ​​of the two Pearson correlation coefficients is used as the exponent of an exponential function with a natural constant as the base, and the product of the calculation result of the exponential function, the temperature mutation degree of the temperature detection sequence, and the sudden decrease index of the oxygen concentration detection sequence is used as the corresponding fluctuation index of the subsequence with the largest trend difference.

[0030] Preferably, the method of completing the abnormality detection of the data collected by the optical fiber sensor according to the corresponding fluctuation index of the subsequence includes:

[0031] Based on the corresponding fluctuation index of the subsequence, the global fluctuation factor of all data collected by the optical fiber sensor is determined, and the global fluctuation factor is used as the neighborhood radius of the DBSCAN clustering algorithm. The normalized values ​​of all data collected by the optical fiber sensor are clustered using the DBSCAN clustering algorithm to obtain clusters, and all data in the clusters with a data quantity less than a preset threshold are regarded as abnormal data.

[0032] Preferably, the method for determining the global fluctuation factor of all data collected by the optical fiber sensor is:

[0033] The information entropy of all collected temperature data and the cumulative value of the information entropy of all collected oxygen concentration data are calculated, and the normalized value of the product of the cumulative value and the corresponding fluctuation index of the subsequence is used as the global fluctuation factor of all data collected by the optical fiber sensor.

[0034] The present invention has at least the following beneficial effects:

[0035] The present invention mainly uses an optical fiber sensor to collect smoke concentration, temperature, and oxygen concentration at each moment, extracts each subsequence in the smoke concentration sequence, obtains the trend difference of each subsequence according to the difference between each subsequence and other subsequences, reflects the difference between each subsequence and other subsequences in the smoke concentration sequence, and improves the reliability of abnormal data recognition in the subsequence; obtains the subsequence with the largest trend difference, uses the subsequence with the largest trend difference to obtain the temperature mutation degree of the temperature data at the corresponding moment, reflects the abnormal degree of the temperature data at the corresponding moment of the subsequence, combines the correlation between smoke concentration and temperature data, and improves the accuracy of abnormal temperature data recognition; uses the subsequence with the largest trend difference to obtain the temperature mutation degree of the temperature data at the corresponding moment of the subsequence, reflects the abnormal degree of the temperature data at the corresponding moment of the subsequence, combines the correlation between smoke concentration and temperature data, and improves the accuracy of abnormal temperature data recognition; The sudden decrease index of oxygen concentration at the corresponding moment is obtained by using the subsequence of the corresponding moment, which reflects the possibility of abnormal oxygen concentration data at the corresponding moment of the subsequence, and combines the correlation between smoke concentration and oxygen concentration; the corresponding fluctuation index of the subsequence is obtained by combining the temperature data at the corresponding moment, the degree of temperature mutation of oxygen concentration, the sudden decrease index and the correlation between smoke concentration and temperature and oxygen concentration. The accuracy of abnormal data identification at each moment is improved through the correlation between smoke concentration, temperature and oxygen concentration at the same moment; finally, the corresponding fluctuation index of the subsequence is used to detect anomalies in the data collected by the optical fiber sensor, which reflects the possibility of abnormal conditions in the geological environment at each moment and improves the accuracy of data anomaly detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flowchart of a method for processing optical fiber sensor data for high-precision monitoring provided by an embodiment of the present invention;

[0038] Figure 2 Construct a flow chart for the fiber optic sensor abnormal data detection indicator. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a high-precision monitoring optical fiber sensor data processing method proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0041] The specific scheme of the optical fiber sensor data processing method for high-precision monitoring provided by the present invention is described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1 , which shows a flowchart of a method for processing optical fiber sensor data for high-precision monitoring provided by an embodiment of the present invention, the method comprising the following steps:

[0043] Step S001, using optical fiber sensors to collect geological data sequences at each moment and perform preprocessing.

[0044] As an embodiment of the present application, optical fiber smoke sensors, optical fiber temperature sensors and optical fiber oxygen sensors are deployed along the transport lanes to continuously monitor the smoke, temperature and oxygen conditions in the lanes in real time. The concentration units of smoke and oxygen are mg / m 3. When using optical fiber sensors to collect tunnel data, the collection time interval is S seconds, and the length of each collected data is T. In this embodiment, S=1, T=36000. The implementer can set it according to the actual situation, and this embodiment does not limit this. Arrange all the collected smoke concentration data, temperature data, and oxygen concentration data in chronological order as smoke concentration sequence, temperature data sequence, and oxygen concentration sequence, respectively, and record them as geological data sequence. In order to avoid the influence of different data dimensions in the smoke concentration sequence, temperature data sequence, and oxygen concentration sequence, the smoke concentration sequence, temperature data sequence, and oxygen concentration sequence are normalized to the maximum value to facilitate subsequent analysis. Maximum value normalization is an existing well-known technology and will not be described in detail in this embodiment. Use the normalized smoke concentration sequence, temperature data sequence, and oxygen concentration sequence to construct a geological anomaly detection matrix K.

[0045]

[0046] All elements in the first row of the geological anomaly detection matrix K are the normalized smoke concentration sequence, denoted as K1, all elements in the second row are the normalized temperature data sequence, denoted as K2, and all elements in the third row are the normalized oxygen concentration sequence, denoted as K3.

[0047] Step S002, obtaining each subsequence of the smoke concentration sequence, calculating the trend difference of each subsequence; using the trend difference of the subsequence and the distribution of the temperature data at the corresponding moment of the subsequence to obtain the temperature mutation degree of the temperature data at the corresponding moment of the subsequence.

[0048] Specifically, this embodiment uses an optical fiber sensor to collect smoke concentration, temperature, and oxygen concentration at each moment, extracts each subsequence of the smoke concentration sequence according to the distribution relationship of the elements in the smoke concentration sequence, obtains the trend difference of any subsequence according to the difference between any subsequence and other subsequences, obtains the temperature mutation degree of the temperature data at the corresponding moment and the sudden decrease index of the oxygen concentration at the corresponding moment according to the trend difference of the subsequence, obtains the corresponding fluctuation index of the subsequence by combining the temperature mutation degree and the sudden decrease index, and completes the abnormal detection of the optical fiber sensor collected data according to the corresponding fluctuation index of the subsequence. The optical fiber sensor abnormal data detection index construction flow chart is shown as follows: Figure 2 The construction process of the corresponding volatility index of the subsequence is as follows:

[0049] Under normal circumstances, the smoke data and temperature data of the transport tunnel also have certain characteristic rules. For example, the smoke data will change greatly during loading and unloading operations, and will also change dramatically with the occurrence of fire; the temperature data will gradually change with the time of the day, and will also change dramatically with the occurrence of fire; the oxygen data is relatively stable under normal circumstances, but will also change greatly with the occurrence of fire.

[0050] By analyzing the smoke data and loading and unloading operations in the transport tunnel, it is found that when loading and unloading minerals in the transport tunnel, a large amount of dust will be generated during the collision, tumbling and crushing of materials such as ore or coal. Driven by the wind, the dust can easily spread throughout the tunnel. In order to ensure the safety of loading and unloading operations, loading and unloading operations must follow strict safety operating procedures, especially in cases where dust, toxic and harmful gases are easily generated. Therefore, the dust concentration has a specific variation law during each loading and unloading operation, and there is a certain similarity between the smoke data, showing a certain degree of periodicity. If smoldering occurs, that is, the previous stage of the open flame phenomenon, the rising law of the smoke will be quite different from the variation law during loading and unloading operations. The rising degree is slow at the beginning, but it will rise sharply afterwards, and the smoke concentration caused by the smoldering phenomenon is much greater than the smoke concentration during loading and unloading operations. At the same time, there is a correlation between smoke data and temperature data. The collision of materials during loading and unloading operations will only lead to an increase in smoke concentration, but not a rise in temperature; and when smoldering, the temperature data will also rise sharply, and the two will have a certain correlation.

[0051] In the smoke concentration sequence K1, the element with the most occurrences is recorded as MS; at the same time, all the maximum values ​​in the smoke concentration sequence are obtained, that is, the element at the position of the maximum value is greater than both the previous element and the next element. Then extract the subsequence from the smoke concentration sequence K1, and the extraction rule is as follows: take the maximum value as the last element of the subsequence, and add the elements in the smoke concentration sequence K1 forward in sequence until MS is the first element of the subsequence. If the MS element cannot be found forward, the subsequence extraction operation is not performed, and each subsequence is obtained. Under normal circumstances, the smoke concentration data fluctuation in the transport lane is relatively stable, while during loading and unloading operations and smoldering, the smoke concentration will change greatly and the data value will become larger, so the value of MS is smaller, and the data in each extracted subsequence will show an overall increasing trend.

[0052] Taking the ith subsequence as an example, the element value in the subsequence is used as the ordinate, and the sequence number of the element value in the subsequence is used as the abscissa. The subsequence is fitted into a straight line using the least squares method, and the slope of the straight line is recorded as k. i ; At the same time, calculate the coefficient of variation of the i-th subsequence and record the result as b i. The least square method and the coefficient of variation are both existing well-known technologies, and this embodiment will not be repeated here. Secondly, as an embodiment of the present application, a resistive strain gauge is used to measure the stress of the ground of the transport tunnel at the same time as the smoke data, and the information entropy and kurtosis of the stress of the ground of the transport tunnel at the corresponding time of all elements in the i-th subsequence are calculated, and the normalized value of the product of the information entropy and the kurtosis is used as the stress mutation factor of the i-th subsequence. During the loading and unloading operation, the stress of the ground of the transport tunnel will also change regularly. If the smoke data is abnormal, it may be due to the irregular loading and unloading operation, and the collision of materials causes the stress of the ground of the transport tunnel to mutate. The larger the information entropy, the more chaotic the stress distribution of the ground of the transport tunnel at the corresponding time of all elements in the i-th subsequence, the greater the possibility of a sudden change in the stress of the ground of the transport tunnel, and the greater the stress mutation factor; the larger the kurtosis, the more extreme values ​​of the stress of the ground of the transport tunnel at the corresponding time of all elements in the i-th subsequence, and the higher the possibility of abnormal values, so the stress mutation factor is also larger. It should be noted that the implementer may choose other methods to measure the stress of the ground in the transport tunnel at the same time as the smoke data according to the actual situation, and this application does not limit this.

[0053] Since the smoke data has a gradual change rule during loading and unloading operations, and each loading and unloading operation must follow a strict operating process, the smoke data has a certain periodicity. Therefore, the trend changes between subsequences during loading and unloading operations are relatively consistent. Therefore, the trend difference of the i-th subsequence is calculated, and the expression is:

[0054]

[0055] In the formula, Q i is the trend difference of the ith subsequence, k i is the slope of the fitted line of the ith subsequence, k j is the slope of the fitted line of the jth subsequence, b i is the coefficient of variation of the ith subsequence, b j is the coefficient of variation of the j-th subsequence, DTW(i,j) is the DTW distance between the i-th subsequence and the j-th subsequence, p is the number of subsequences of the smoke concentration sequence K1, δ i is the stress mutation factor of the stress of the ground surface of the transport tunnel at the corresponding moment of all elements in the i-th subsequence. The DTW distance is a well-known technology, and will not be described in detail in this embodiment.

[0056] When Q i The smaller it is, the higher the similarity between the ith subsequence and all other subsequences is, that is, the smaller the trend difference of the ith subsequence is, the more similar the changes of the ith subsequence and other subsequences are, and the more likely the ith subsequence is normal smoke data, the more likely it is that there is no abnormality.

[0057] Get the subsequence r with the largest trend difference, and let the length of the subsequence r with the largest trend difference be v. Then record the collection time interval of the smoke data of the subsequence with the largest trend as [u,u+v-1]. At the same time, analyze the change of temperature data in the time interval [u,u+v-1] to determine whether a sharp increase has occurred. The temperature data in the time interval [u,u+v-1] is recorded as the temperature detection sequence c2. The temperature will change with the changes at different times of the day. The data values ​​in the morning, evening and night are smaller, and the data values ​​at noon and daytime are larger. Although there is no specific pattern in the change, the stability is relatively high, and multiple fluctuations will not occur in a short period of time. Therefore, the degree of temperature mutation of the temperature detection sequence c2 is constructed, recorded as W c2 , the expression is:

[0058] G c2 =|K2 u+v―1 ―K2 u |+|K2 u+v―1 ―max(K2 u,u+v―1 )|

[0059]

[0060] In the formula, G c2 is the maximum score of the temperature detection sequence c2, K2 u+v―1 Represents the element at the u+v-1th moment of the temperature data sequence K2, K2 u Represents the element at the uth moment of the temperature data sequence K2, K2 u,u+v―1 represents all elements of the temperature data sequence K2 from the uth moment to the u+v-1th moment, max() represents the maximum value function; | K2 u+v―1 ―K2 u | is recorded as the absolute value of the first difference, |K2 u+v―1 ―max(K2 u,u+v―1 )| is recorded as the absolute value of the second difference;

[0061] W c2 is the temperature mutation degree of temperature detection sequence c2, std c2 is the standard deviation of all elements of the temperature detection sequence, is the standard deviation of all collected temperature data, δ is a parameter adjustment coefficient preset to be greater than 0, c2 i+1 、c2 i The i+1th and ith elements of the temperature detection sequence are respectively, and v is the number of elements in the temperature detection sequence. In this embodiment, δ=0.01, and the implementer can set it according to the actual situation, and this embodiment does not limit it here.

[0062] Under normal circumstances, according to the law of stable temperature change, we can get: |K2u+v―1 ―K2 u | has a smaller value, K2 u+v―1 ―max(K2 u,u+v―1 ) is also small; and because the temperature changes steadily, there will not be large fluctuations in a short period of time, std c2 and The sizes of are similar and the values ​​are small. At the same time, since the temperature is stable, the sizes of adjacent elements in the temperature detection sequence c2 are close. If smoldering occurs, the temperature will rise sharply and exceed the highest temperature in history. At this time: |K2 u+v―1 ―K2 u | has a larger value; K2 u+v―1 ―max(K2 u,u+v―1 ) is also large; and due to the rapid increase in temperature, std c2 The value of will be larger, and due to It still meets the characteristics of stable temperature change, and its value is small, so The value of will be larger; at the same time, the temperature increases sharply, then The value of is also larger.

[0063] At this point, by calculating the temperature mutation degree W of the temperature detection sequence c2 c2 , we can judge whether the temperature change is significantly different from the previous temperature data and whether it conforms to the characteristics of stable temperature change. c2 The smaller the value of is, the more it conforms to the characteristics of stable temperature change; if W c2 The larger the value, the more rapidly the temperature is rising and the greater the possibility of an impending fire.

[0064] Step S003, obtaining a sudden decrease index of the oxygen concentration data at the corresponding moment of the subsequence according to the trend difference of the subsequence and the distribution of the oxygen concentration data at the corresponding moment of the subsequence.

[0065] In the transport tunnel under normal conditions, the oxygen content is always relatively consistent. If smoldering occurs, the oxygen content will decrease sharply and continue to decrease. The oxygen concentration data in the time interval [u,u+v-1] is recorded as the oxygen concentration detection sequence c3, and the sudden decrease index of the oxygen concentration detection sequence c3 is calculated and recorded as Z c3 , the expression is:

[0066] num=Num{c3 i+1 ―c3 i ≤0},i∈[u,u+v―2]

[0067]

[0068] Where num is the decreasing number of oxygen concentration detection sequence c3, Num{} is a statistical function, which counts the number of values ​​less than or equal to 0, c3 i 、c3 i+1 are the i-th and i+1-th elements in the oxygen concentration detection sequence c3 respectively; i+1 ―c3 i Recorded as the first difference;

[0069] Z c3 is the sudden decrease index of the oxygen concentration detection sequence c3, f is the number of oxygen concentration data that is less than the mean of all collected oxygen concentration data, T is the number of all collected oxygen concentration data, is the mean value of all elements in the oxygen concentration detection sequence c3, is the average value of all oxygen concentration data collected. Recorded as the absolute value of the third difference.

[0070] At this point, by calculating the sudden decrease degree Z of the oxygen concentration detection sequence c3 c3 , we can calculate and make a more accurate judgment on the oxygen situation inside the transport tunnel. c3 The smaller the value, the normal the oxygen situation. c3 The larger the value, the greater the possibility of fire.

[0071] Step S004, combining the temperature data at the corresponding time, the temperature mutation degree of the oxygen concentration, the sudden drop index, and the correlation between the smoke concentration and the temperature and oxygen concentration to obtain the corresponding fluctuation index of the subsequence.

[0072] Combining the temperature data, the temperature mutation degree of oxygen concentration, the sudden drop index, and the correlation between smoke concentration and temperature and oxygen concentration, the corresponding fluctuation index of the subsequence with the largest trend difference is calculated, which is recorded as BD u , the expression is:

[0073]

[0074] In the formula, BD r is the corresponding volatility index of the subsequence r with the largest trend difference, |P rc2 |、|P rc3 | are the absolute values ​​of the Pearson correlation coefficients between the subsequence r and the temperature detection sequence c2 and the oxygen concentration detection sequence c3, respectively. c3 is the sudden decrease index of oxygen concentration detection sequence c3, W c2 is the temperature mutation degree of temperature detection sequence c2, and e is a natural constant.

[0075] Under normal circumstances, |P rc2 |with|P rc3The values ​​of | are all small. c2 is the temperature mutation coefficient of temperature detection sequence c2, W c2 The smaller the value, the smaller the temperature fluctuation, which is more consistent with the temperature change characteristics under normal conditions. c3 is the sudden decrease degree of oxygen concentration detection sequence, Z c3 The smaller the value, the more it conforms to the changing characteristics of oxygen under normal conditions. r The smaller the value is, the more normal the subsequence r and the corresponding temperature data and oxygen data are, and the smaller the possibility of smoldering at the corresponding moment of the subsequence r is.

[0076] Step S005 , performing anomaly detection on the data collected by the optical fiber sensor according to the corresponding fluctuation index of the subsequence with the largest trend difference.

[0077] Using the corresponding fluctuation index of the subsequence r with the largest trend difference, combined with temperature data and oxygen concentration data, the global fluctuation factor of all data collected by the optical fiber sensor is calculated. The expression is:

[0078] HZ K =norm{BD r ×(Ent K2 +Ent K3 )}

[0079] In the formula, HZ K It represents the global fluctuation factor of all data collected by the optical fiber sensor, norm is the normalization function, BD r is the corresponding volatility index of the subsequence r with the largest trend difference, Ent K2 is the information entropy of all elements in the collected temperature data sequence K2, Ent K3 is the information entropy of all elements of the collected oxygen concentration sequence K3.

[0080] The normalized value of all data collected by the optical fiber sensor, that is, all data in the anomaly detection matrix, is used as the input of the DBSCAN clustering algorithm, and the global fluctuation factor of all data collected by the optical fiber sensor is used as the neighborhood radius of the DBSCAN clustering algorithm. The value of the minimum sample point MinPts in this embodiment is 20, which can be set by the implementer according to the actual situation. The output of the DBSCAN clustering algorithm is each cluster cluster. All cluster clusters containing data less than τ in the cluster cluster are counted and marked. All data in the marked cluster cluster are regarded as abnormal data, indicating that the data collected at the corresponding time may have errors, or there may be fire risks at the corresponding time, and data anomalies occur. In this embodiment, τ=25, which can be set by the implementer according to the actual situation. This embodiment is not limited here. It should be noted that the use of the global fluctuation factor as the neighborhood radius of the DBSCAN clustering algorithm combines the correlation between smoke data, oxygen concentration data, and temperature data, as well as the change and distribution trend, takes into account the inherent structural distribution characteristics of the data, improves the accuracy of the neighborhood radius determination, and thus improves the accuracy of abnormal data identification of the optical fiber sensor.

[0081] In summary, the embodiment of the present invention uses an optical fiber sensor to collect smoke concentration, temperature, and oxygen concentration at each moment, extracts each subsequence in the smoke concentration sequence, obtains the trend difference of each subsequence according to the difference between each subsequence and other subsequences, reflects the difference between each subsequence and other subsequences in the smoke concentration sequence, and improves the reliability of abnormal data identification in the subsequence; obtains the subsequence with the largest trend difference, and uses the subsequence with the largest trend difference to obtain the temperature mutation degree of the temperature data at the corresponding moment, reflects the abnormal degree of the temperature data at the corresponding moment of the subsequence, combines the correlation between smoke concentration and temperature data, and improves the accuracy of abnormal temperature data identification; uses the trend difference The largest subsequence obtains the sudden decrease index of oxygen concentration at the corresponding moment, which reflects the possibility of abnormal oxygen concentration data at the corresponding moment of the subsequence, combined with the correlation between smoke concentration and oxygen concentration; the corresponding fluctuation index of the subsequence is obtained by combining the temperature data at the corresponding moment, the degree of temperature mutation of oxygen concentration, the sudden decrease index and the correlation between smoke concentration and temperature, oxygen concentration. Through the correlation between smoke concentration, temperature and oxygen concentration at the same moment, the accuracy of abnormal data identification at each moment is improved; finally, the corresponding fluctuation index of the subsequence is used to detect anomalies in the data collected by the optical fiber sensor, which reflects the possibility of abnormal conditions in the geological environment at each moment and improves the accuracy of data anomaly detection.

[0082] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for processing optical fiber sensor data for high-precision monitoring, characterized in that: The method comprises the following steps: Use optical fiber sensors to collect smoke concentration, temperature, and oxygen concentration at each moment; The smoke concentrations at all acquisition moments are combined into a smoke concentration sequence, and each subsequence of the smoke concentration sequence is extracted according to the distribution relationship of the elements in the smoke concentration sequence; the trend difference of any subsequence is obtained according to the difference between any subsequence and other subsequences; According to the trend difference of the subsequences and the distribution of the temperature data at the corresponding time, the temperature mutation degree of the temperature data at the corresponding time is obtained; according to the trend difference of the subsequences and the distribution of the oxygen concentration at the corresponding time, the sudden decrease index of the oxygen concentration at the corresponding time is obtained; The corresponding fluctuation index of the subsequence is obtained by combining the temperature data at the corresponding moment, the temperature mutation degree of oxygen concentration, the sudden drop index, and the correlation between smoke concentration and temperature and oxygen concentration; The abnormality detection of the data collected by the optical fiber sensor is completed according to the corresponding fluctuation index of the subsequence; The temperature mutation degree of the temperature data at the corresponding moment includes: All temperature data at the corresponding time of the subsequence with the largest trend difference are combined into a temperature detection sequence, and the maximum score of the temperature detection sequence is determined based on the degree of difference between the elements in the temperature detection sequence; The temperature mutation degree of temperature detection sequence c2 is recorded as W c2 , the expression is: In the formula, G c2 is the maximum score of the temperature detection sequence c2, std c2 is the standard deviation of all elements of the temperature detection sequence, is the standard deviation of all collected temperature data, δ is a parameter adjustment coefficient preset to be greater than 0, c2 i+1 、c2 i are the i+1th and ith elements of the temperature detection sequence respectively, and v is the number of elements contained in the temperature detection sequence; The method for determining the maximum score of the temperature detection sequence is: The absolute value of the difference between the first element and the last element in the temperature detection sequence is calculated, recorded as the first difference absolute value, the absolute value of the difference between the last element in the temperature detection sequence and the maximum value in the temperature detection sequence is calculated, recorded as the second difference absolute value, and the sum of the first difference absolute value and the second difference absolute value is used as the maximum value score of the temperature detection sequence.

2. The method for processing optical fiber sensor data for high-precision monitoring according to claim 1, characterized in that: The step of extracting subsequences of the smoke density sequence according to the distribution relationship of elements in the smoke density sequence includes: For the smoke concentration sequence, the element with the most occurrences is recorded as MS, all the maximum values ​​in the smoke concentration sequence are obtained, and any maximum value is used as the tail element of the subsequence. The elements in the smoke concentration sequence are obtained in sequence from any maximum value as the elements in the subsequence until the first MS element is traversed to the first element of the subsequence.

3. The method for processing optical fiber sensor data for high-precision monitoring according to claim 1, characterized in that: The trend difference of any subsequence includes: Fit all elements of each subsequence into straight lines, obtain the slope of each straight line, calculate the absolute value of the difference between the slope of the fitted straight line of any subsequence and other subsequences, calculate the absolute value of the difference between the coefficient of variation of any subsequence and other subsequences, calculate the DTW distance between any subsequence and other subsequences, calculate the product of two absolute values ​​of the difference and the DTW distance, calculate the mean of the product between any subsequence and all other subsequences, calculate the information entropy and kurtosis of the stress of the transport tunnel surface at the corresponding time of all elements in any subsequence, use the multiplication result of the information entropy and the kurtosis as the stress mutation factor, and use the product of the stress mutation factor and the mean as the trend difference of any subsequence.

4. The method for processing optical fiber sensor data for high-precision monitoring according to claim 1, characterized in that: The sudden decrease index of oxygen concentration at the corresponding moment includes: All oxygen concentration data at the corresponding time of the subsequence with the largest trend difference are combined into an oxygen concentration detection sequence, and the decreasing number of the oxygen concentration detection sequence is determined based on the difference between adjacent elements in the oxygen concentration detection sequence; Calculate the mean of all the collected oxygen concentration data, count the number of oxygen concentration data less than the mean, calculate the absolute value of the difference between the mean and the mean of all elements in the oxygen concentration detection sequence, record it as the third absolute value of the difference, calculate the ratio of the number of all the collected oxygen concentration data to the number, and use the product of the ratio, the third absolute value of the difference and the number of decreases as the sudden decrease index of the oxygen concentration detection sequence.

5. The method for processing optical fiber sensor data for high-precision monitoring according to claim 4, characterized in that: The method for determining the decreasing number of times of the oxygen concentration detection sequence is as follows: The difference between each element and the previous element in the oxygen concentration detection sequence is calculated and recorded as a first difference, and the number of all the first differences that are less than or equal to 0 is used as the decreasing number of the oxygen concentration detection sequence.

6. The method for processing optical fiber sensor data for high-precision monitoring according to claim 4, characterized in that: The corresponding volatility index of the subsequence includes: The Pearson correlation coefficients of the subsequence with the largest trend difference and the temperature detection sequence and the oxygen concentration detection sequence are calculated respectively, and the sum of the absolute values ​​of the two Pearson correlation coefficients is used as the exponent of an exponential function with a natural constant as the base, and the product of the calculation result of the exponential function, the temperature mutation degree of the temperature detection sequence, and the sudden decrease index of the oxygen concentration detection sequence is used as the corresponding fluctuation index of the subsequence with the largest trend difference.

7. The method for processing optical fiber sensor data for high-precision monitoring according to claim 1, characterized in that: The method of completing abnormality detection of data collected by the optical fiber sensor according to the corresponding fluctuation index of the subsequence includes: Based on the corresponding fluctuation index of the subsequence, the global fluctuation factor of all data collected by the optical fiber sensor is determined, and the global fluctuation factor is used as the neighborhood radius of the DBSCAN clustering algorithm. The normalized values ​​of all data collected by the optical fiber sensor are clustered using the DBSCAN clustering algorithm to obtain clusters, and all data in the clusters with a data quantity less than a preset threshold are regarded as abnormal data.

8. The method for processing optical fiber sensor data for high-precision monitoring according to claim 7, characterized in that: The method for determining the global fluctuation factor of all data collected by the optical fiber sensor is: The information entropy of all collected temperature data and the cumulative value of the information entropy of all collected oxygen concentration data are calculated, and the normalized value of the product of the cumulative value and the corresponding fluctuation index of the subsequence is used as the global fluctuation factor of all data collected by the optical fiber sensor.

Images