An abnormal state dynamic diagnosis system and method for a gas leakage monitoring device

Abstract

The application discloses a kind of abnormal state dynamic diagnosis system and method of gas leakage monitoring equipment, system includes monitoring data acquisition unit, basic data acquisition unit, data processing unit and abnormal diagnosis unit.The method includes: clear gas leakage monitoring equipment and monitoring gas type, obtain the concentration data, wind direction data and time data of monitoring equipment, in combination with equipment detection limit and normal concentration of monitoring gas in ambient air, establish the sequence to be diagnosed monitoring data;The slope coefficient sequence of concentration data, the trend coefficient sequence of concentration data, and the concentration time distribution sequence and its dispersion degree, the concentration wind direction distribution sequence and its dispersion degree are calculated;According to each sequence calculated, complete abnormal diagnosis.The system and method disclosed in the application are based on real-time pollution monitoring data and real-time state parameter monitoring data, and the abnormal state dynamic diagnosis of monitoring equipment is carried out in combination with equipment performance index, and the monitoring is sensitive, real-time, accurate.

Description

Technical Field

[0001] This invention relates to the field of environmental pollution monitoring, and in particular to a dynamic diagnostic system and method for abnormal states of gas leak monitoring equipment. Background Technology

[0002] With the comprehensive management of volatile organic compounds in key industries and the construction of an environmental risk early warning system for toxic and harmful gases, chemical industrial parks and petrochemical enterprises have gradually built gas pollution monitoring networks by deploying online environmental monitoring equipment, thus realizing online monitoring of pollutant concentrations.

[0003] However, as the lifespan of equipment increases, problems such as sampling pipeline contamination, sensor failure, and circuit malfunctions inevitably lead to a decrease in equipment performance, resulting in abnormal monitoring data. Currently, most equipment faults are diagnosed based on operating status parameters. For example, abnormal sampling flow, temperature, or pressure indicate that the equipment is in an abnormal state. Since small equipment does not have the ability to monitor status parameters, and abnormal monitoring data caused by sensor failure, sampling pipeline contamination, etc., is difficult to diagnose through monitoring equipment operating status parameters, the diagnostic methods based on status parameters have a relatively narrow range of applications.

[0004] Therefore, it is necessary to develop a method for diagnosing the operational status of gas leak monitoring equipment based on monitoring data, so as to diagnose and identify whether the equipment is operating normally and effectively online. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a dynamic diagnostic system and method for abnormal states of gas leak monitoring equipment. Based on real-time pollution monitoring data and real-time status parameter monitoring data, and combined with equipment performance indicators, the system performs dynamic diagnostics of abnormal states of the monitoring equipment to accurately and promptly identify the current operating status of the equipment.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A dynamic diagnostic system for abnormal states of a gas leak monitoring device includes a monitoring data acquisition unit, a basic data acquisition unit, a data processing unit, and an anomaly diagnosis unit. The monitoring data acquisition unit and the basic data acquisition unit respectively input the acquired data into the data processing unit, and the data processing unit inputs the processed data into the anomaly diagnosis unit.

[0008] In the above scheme, the monitoring data acquisition unit includes a concentration data acquisition module, a wind direction data acquisition module, and a time data acquisition module.

[0009] In the above scheme, the basic data acquisition unit includes a device detection limit acquisition module and an ambient air normal concentration acquisition module.

[0010] In the above scheme, the data processing unit includes a slope coefficient sequence processing module, a trend coefficient sequence processing module, a concentration time distribution sequence and its dispersion processing module, and a concentration wind direction distribution sequence and its dispersion processing module.

[0011] A method for dynamic diagnosis of abnormal states of a gas leak monitoring device includes the following steps:

[0012] Step 1: Identify the type of gas being monitored by the gas leak monitoring equipment, obtain the concentration data, wind direction data, and time data of the monitoring equipment, and establish a monitoring data sequence to be diagnosed by combining the equipment detection limit and the normal concentration of the monitored gas in ambient air.

[0013] Step 2: Based on the monitoring data sequence to be diagnosed, calculate the slope coefficient sequence of the concentration data, the trend coefficient sequence of the concentration data, the concentration time distribution sequence of the concentration data on the time data and its dispersion, and the concentration wind direction distribution sequence of the concentration data on the wind direction data and its dispersion.

[0014] Step 3: Complete the anomaly diagnosis based on the calculated sequences and whether there are pollution sources in the surrounding area.

[0015] In the above scheme, the method for constructing the monitoring data sequence to be diagnosed is as follows:

[0016] The monitoring data sequence to be diagnosed includes concentration data, wind direction data, and time data when the monitoring data was acquired.

[0017] If the normal concentration is greater than 3 times the equipment detection limit, then a diagnostic monitoring data sequence containing three-dimensional parameters including time data, wind direction data, and concentration data will be directly constructed.

[0018] Otherwise, select concentration data with values ​​greater than the normal concentration, and combine them with wind direction data and time data at the same time as the concentration data to construct a monitoring data sequence to be diagnosed.

[0019] In the above scheme, the method for establishing the slope coefficient sequence is as follows:

[0020] Slope coefficient sequence k i The calculation formula is as follows:

[0021]

[0022] Among them, c i Let c be the i-th concentration data in the monitoring data sequence to be diagnosed. i-1 C0 represents the (i-1)th concentration data in the monitoring data sequence to be diagnosed, where C0 = C1; C_a represents the average concentration data in the monitoring data sequence to be diagnosed; and M represents the number of concentration data in the monitoring data sequence to be diagnosed.

[0023] In the above scheme, the method for establishing the trend coefficient sequence is as follows:

[0024] First, calculate the moving average data f of the concentration data from the j-th to the i-th data points in the monitoring data sequence to be diagnosed. i When i > H, j∈[1, M]; when H≥i≥1, f i =c i f0 = f1; H is the length of the moving window; c j This refers to the j-th concentration data in the monitoring data sequence to be diagnosed;

[0025] Then, calculate the trend coefficient sequence g. i :

[0026]

[0027] Among them, f i-1 C_a represents the moving average of the concentration data from the (j-1)th to the (i-1)th concentration data in the monitoring data sequence to be diagnosed; C_a represents the average concentration data in the monitoring data sequence to be diagnosed; and M represents the number of concentration data in the monitoring data sequence to be diagnosed.

[0028] In the above scheme, the method for calculating the concentration time distribution sequence and its dispersion of concentration data over time is as follows:

[0029] First, collect the concentration data of the monitoring device to be diagnosed over the past few days and establish a monitoring data sequence to be diagnosed;

[0030] Then, the 0-24 hour period is divided into 24 or 12 time periods. Based on the time data in the monitoring data sequence to be diagnosed, the collected concentration data is divided into each time period, and the average value of multiple concentration data in each time period is calculated.

[0031] Finally, the dispersion s of the average concentration data within each time period is calculated:

[0032]

[0033] Among them, t j Let be the average concentration in the j-th time period, where j∈[1, p], and p is the number of time periods. To substitute all t into the calculation j The average value.

[0034] In the above scheme, the method for calculating the concentration-wind-direction distribution sequence of concentration data and its dispersion in the wind-direction data is as follows:

[0035] First, collect the concentration data of the monitoring device to be diagnosed over the past few days and establish a monitoring data sequence to be diagnosed;

[0036] Then, the 0-360° wind direction is divided into 8, 12 or 16 wind direction intervals. Based on the wind direction data in the monitoring data sequence to be diagnosed, the collected concentration data is divided into each wind direction interval, and the average value of multiple sets of concentration data in each interval is calculated.

[0037] Finally, the dispersion r of the average concentration data within each wind direction interval is calculated:

[0038]

[0039] Among them, w v Let be the average concentration in the v-th wind direction interval, where v∈[1,q], and q is the number of wind direction intervals. To substitute all w into the calculation v The average value.

[0040] In the above scheme, step three is performed as follows:

[0041] First, perform anomaly diagnosis on the slope coefficient sequence and trend coefficient sequence respectively. If either result is abnormal, the monitoring equipment is considered abnormal and the process ends. Otherwise, determine whether there is a pollution source around the monitoring equipment. If not, the monitoring equipment is considered normal and the process ends. If so, perform anomaly diagnosis on the concentration time distribution sequence and its dispersion, and the concentration wind direction distribution sequence and its dispersion respectively. If either result is abnormal, the monitoring equipment is considered abnormal and the process ends. Otherwise, the monitoring equipment is considered normal and the process ends.

[0042] In a further technical solution, the specific method for step three is as follows:

[0043] Step 1: If in the slope coefficient sequence, k i The percentage of data with a value ≥0 exceeds the threshold A or k. i If the proportion of data with a value ≤0 exceeds the threshold A, it is judged as abnormal and the process ends; otherwise, proceed to step 2.

[0044] Step 2: If in the trend coefficient sequence, g i The percentage of data with a value ≥0 exceeds the threshold B or g. i If the proportion of data with a value ≤0 exceeds the threshold B, it is judged as abnormal and the process ends; otherwise, proceed to step 3.

[0045] Step 3: Determine if there is a pollution source in the vicinity. If not, the instrument will function normally and the process will end. If so, proceed to Step 4.

[0046] Step 4: Determine the dispersion of the concentration time distribution sequence. If the obtained dispersion s is less than the threshold... If the concentration is the average value across all time periods, then the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, proceed to step 5.

[0047] Step 5: Determine the dispersion of the concentration-wind direction distribution sequence. If the obtained dispersion r is less than the threshold... If the concentration is the average value across all wind directions, then the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, the monitoring equipment is considered normal, and the diagnosis ends.

[0048] In a further technical solution, the threshold A ranges from 80% to 100%.

[0049] In a further technical solution, the threshold B ranges from 80% to 100%.

[0050] In a further technical solution, the threshold C ranges from 5% to 50%.

[0051] In a further technical solution, the threshold D ranges from 5% to 50%.

[0052] In a further technical solution, the number of M is not less than the number of real-time data obtained by the monitoring device to be diagnosed over 6 hours.

[0053] In a further technical solution, the length H of the moving window is set to 3-6.

[0054] In a further technical solution, the number of concentration data in the monitoring data sequence to be diagnosed is the number of real-time data obtained by the monitoring device to be diagnosed over 3-7 days.

[0055] In the above scheme, if the monitoring equipment can monitor multiple factors simultaneously, each factor needs to be diagnosed separately for abnormalities.

[0056] In the above scheme, if the key status parameters of the monitoring equipment exceed the normal range, it is also considered abnormal. The key status parameters include the pressure, temperature, flow rate, current, and voltage of key components.

[0057] The present invention provides a dynamic diagnostic method for abnormal states of gas leak monitoring equipment, which has the following beneficial effects through the above technical solution:

[0058] This invention calculates the slope coefficient sequence, trend coefficient sequence, concentration time distribution sequence and dispersion of concentration data, and concentration wind direction distribution sequence and dispersion of concentration data based on the monitoring data sequence to be diagnosed. Anomaly diagnosis is completed by using the data sequence obtained from the above calculations and the presence of pollution sources in the surrounding area.

[0059] This invention is based on pollution monitoring data and combines equipment performance indicators with the characteristics of normal pollution monitoring data for diagnosis. This method directly uses pollution monitoring data as the diagnostic object, and uses the slope characteristics, trend characteristics, concentration time distribution characteristics, and concentration wind direction distribution characteristics of the monitoring data as diagnostic criteria to distinguish the differences between normal and abnormal equipment states. It can further discover non-mechanical and electrical structural problems such as sensor failure. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0061] Figure 1 This is a schematic diagram of an abnormal state dynamic diagnosis system for a gas leak monitoring device disclosed in an embodiment of the present invention;

[0062] Figure 2 This is a data relationship diagram on which the embodiments of the present invention are based.

[0063] Figure 3 This is a schematic diagram of a dynamic diagnostic method for abnormal states of a gas leak monitoring device disclosed in an embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0065] This invention provides a dynamic diagnostic system for abnormal states of gas leak monitoring equipment, such as... Figure 1 As shown, it includes a monitoring data acquisition unit, a basic data acquisition unit, a data processing unit, and an anomaly diagnosis unit. The monitoring data acquisition unit and the basic data acquisition unit respectively input the acquired data into the data processing unit, and the data processing unit inputs the processed data into the anomaly diagnosis unit.

[0066] The monitoring data acquisition unit includes a concentration data acquisition module, a wind direction data acquisition module, and a time data acquisition module. The basic data acquisition unit includes an equipment detection limit acquisition module and an ambient air normal concentration acquisition module. The data processing unit includes a slope coefficient sequence processing module, a trend coefficient sequence processing module, a concentration time distribution sequence and its dispersion processing module, and a concentration wind direction distribution sequence and its dispersion processing module.

[0067] Example 1

[0068] This invention provides a method for dynamic diagnosis of abnormal states in gas leak monitoring equipment, such as... Figure 2 As shown, it includes the following steps:

[0069] Step 1: Identify the type of gas being monitored by the gas leak monitoring equipment, obtain the concentration data, wind direction data, and time data from the monitoring equipment, and establish a monitoring data sequence to be diagnosed by combining the equipment's detection limit and the normal concentration of the monitored gas in ambient air.

[0070] 1. Construction of monitoring data sequences to be diagnosed

[0071] Taking an online monitoring device for VOCs in ambient air as an example, with a VOCs detection limit of 5 ppb, a measurement cycle of 1 minute, and a normal concentration of 50 ppb in ambient air, the specific method for constructing the monitoring data sequence to be diagnosed is explained below.

[0072] First, acquire the monitoring data of the monitoring equipment in the recent period, including concentration data, wind direction data, and time data. Specifically, the concentration data refers to the VOCs concentration data obtained from the monitoring.

[0073] Since the normal concentration (50 ppb) is greater than 3 times the detection limit (5 ppb), the monitoring data sequence to be diagnosed can be directly constructed, as shown in Table 1.

[0074] Table 1. Monitoring data sequence to be diagnosed

[0075] Time data Concentration data (ppb) Wind direction data (°) 2020.07.12 11.00 189 312 2020.07.12 11.01 212 301 2020.07.12 11.02 296 330 2020.07.12 11.03 267 320 2020.07.12 11.04 185 321 ... ... ...

[0076] If a VOCs gas detection device has a detection limit of 70 ppb, a measurement cycle of 1 minute, and a normal concentration of 50 ppb in ambient air, when constructing the monitoring data sequence to be diagnosed, it is necessary to select concentration data that are more than 3 times the detection limit, i.e., 210 ppb or higher. Combine the wind direction data and time data at the same time as the concentration data to construct the monitoring data sequence to be diagnosed, as shown in Table 2.

[0077] Table 2. Monitoring data sequence to be diagnosed

[0078] Time data Concentration data (ppb) Wind direction data (°) 2020.07.12 11.01 212 301 2020.07.12 11.02 296 330 2020.07.12 11.03 267 320 ... ... ...

[0079] Step 2: Based on the monitoring data sequence to be diagnosed, calculate the slope coefficient sequence of the concentration data, the trend coefficient sequence of the concentration data, the concentration time distribution sequence and its dispersion of the concentration data on the time data, and the concentration wind direction distribution sequence and its dispersion of the concentration data on the wind direction data.

[0080] 2. Construction of slope coefficient sequence

[0081] The method for establishing the slope coefficient sequence is as follows. Taking an online VOCs monitoring device for ambient air as an example, firstly, the monitoring data of the VOCs monitoring factors of the device for the past 6 hours are obtained, and the monitoring data sequence to be diagnosed is constructed. Example data is shown in Table 3, C1-C 360 The data represents concentration, with an average value of 178.

[0082] Table 3. Monitoring data sequence to be diagnosed

[0083] <![CDATA[C1]]> <![CDATA[C2]]> <![CDATA[C3]]> <![CDATA[C4]]> <![CDATA[C5]]> <![CDATA[C6]]> <![CDATA[C i ]]> <![CDATA[C 258 ]]> <![CDATA[C 259 ]]> <![CDATA[C 360 ]]> 125 134 159 207 184 198 ... 177 152 109

[0084] Then, the slope coefficient k of each concentration data point in the monitoring data sequence to be diagnosed is calculated compared to the previous data point. i ,Right now

[0085]

[0086] Among them, c i Let c be the i-th concentration data in the monitoring data sequence to be diagnosed. i-1 Let C0 be the (i-1)th concentration data point in the monitoring data sequence to be diagnosed, where C0 = C1; C_a is the average concentration data in the monitoring data sequence to be diagnosed; and M is the number of concentration data points in the monitoring data sequence to be diagnosed, which should be no less than the number of real-time data points that the device can obtain in 6 hours. The calculation results are shown in Table 4.

[0087] Table 4 Slope Coefficient Sequence

[0088] <![CDATA[k1]]> <![CDATA[k2]]> <![CDATA[k3]]> <![CDATA[k4]]> <![CDATA[k5]]> <![CDATA[k6]]> <![CDATA[k i ]]> <![CDATA[k 258 ]]> <![CDATA[k 259 ]]> <![CDATA[k 360 ]]> 0 0.05 0.14 0.26 -0.12 0.07 ... ... -0.14 -0.24

[0089] 3. Construction of trend coefficient series

[0090] The method for establishing the trend coefficient sequence is as follows. Taking the online monitoring equipment for ambient air VOCs as an example, firstly, the monitoring data of the VOCs monitoring factors of the equipment for the past 6 hours are obtained, and the monitoring data sequence to be diagnosed is constructed. An example is shown in Table 3.

[0091] Then, calculate the moving average data f of the concentration data from the j-th to the i-th concentration data in the monitoring data sequence to be diagnosed. i When i > H, j∈[1, M]; when H≥i≥1, f i =c i f0 = f1; H is the length of the moving window, recommended value is 3-8, this embodiment uses 3; c j The j-th concentration data in the monitoring data sequence to be diagnosed is shown in Table 5.

[0092] Table 5 Moving Average Data

[0093] <![CDATA[f1]]> <![CDATA[f2]]> <![CDATA[f3]]> <![CDATA[f4]]> <![CDATA[f5]]> <![CDATA[f6]]> <![CDATA[f i ]]> <![CDATA[f 358 ]]> <![CDATA[f 359 ]]> <![CDATA[f 360 ]]> 125 134 159 166.6 183.3 196.3 ... ... ... 146.0

[0094] Finally, the trend coefficient sequence g is calculated. i :

[0095]

[0096] Among them, f i-1C_a represents the moving average of the concentration data from the (j-1)th to the (i-1)th concentration data in the monitoring data sequence to be diagnosed; C_a represents the average concentration data in the monitoring data sequence to be diagnosed; M represents the number of concentration data in the monitoring data sequence to be diagnosed, which should be no less than the number of real-time data that the device can obtain in 6 hours. If the instrument response cycle is long and the number of data sets acquired is less than 300, the data collection time can be extended. The calculation results are shown in Table 6.

[0097] Table 6 Trend Coefficient Series

[0098] <![CDATA[g1]]> <![CDATA[g2]]> <![CDATA[g3]]> <![CDATA[g4]]> <![CDATA[g5]]> <![CDATA[g6]]> <![CDATA[g i ]]> <![CDATA[G 358 ]]> <![CDATA[G 359 ]]> <![CDATA[g 360 ]]> 0 0.05 0.14 0.04 0.09 0.07 ... ... ... ...

[0099] 4. Construction of concentration time distribution series

[0100] First, the concentration data of the monitoring device to be diagnosed was collected for the past 3 days (assuming data is transmitted once per minute, a total of 4320 sets of monitoring data were collected over 3 days), and the monitoring data sequence to be diagnosed was established. The results are shown in Table 1.

[0101] Then, the 0-24 hour period was divided into 12 time periods. Based on the time data in the monitoring data sequence to be diagnosed, the 4320 sets of monitoring data were divided into each time period, and the average concentration data of multiple sets of monitoring data within the time period was calculated. The results are shown in Table 7.

[0102] Table 7 Concentration Time Distribution Series

[0103] Time period Number of monitoring data groups included in this segment Average concentration data 0:00-2:00 360 197 2 o'clock - 4 o'clock 360 215 4 o'clock - 6 o'clock 360 245 6 o'clock - 8 o'clock 360 256 8:00-10:00 360 198 10:00-12:00 360 154 12:00-14:00 360 132 2 PM - 4 PM 360 149 4 PM - 6 PM 360 178 6 PM - 8 PM 360 219 8 PM - 10 PM 360 236 10 PM - 12 AM 360 215

[0104] Finally, the dispersion s of the average concentration data within each time period was calculated: the data are shown in Table 8.

[0105]

[0106] Among them, t j Let be the average concentration in the j-th time period, where j∈[1, p], and p is the number of time periods. To substitute all t into the calculation j The average value. The recommended number of concentration data points for the diagnostic monitoring data sequence is the number of real-time data points available from the device over 3-7 days.

[0107] Table 8. Average concentrations for each time period

[0108]

[0109]

[0110] 5. Construction of Concentration-Wind Direction Distribution Sequence

[0111] First, the concentration data of the monitoring device to be diagnosed was collected for the past 3 days (assuming data is transmitted once per minute, a total of 4320 sets of monitoring data were collected over 3 days), and the monitoring data sequence to be diagnosed was established. The results are shown in Table 1.

[0112] Then, the 0-360° wind direction was divided into 12 wind direction intervals. Based on the wind direction data in the monitoring data sequence to be diagnosed, 4320 sets of monitoring data were divided into each wind direction interval, and the average concentration data of multiple sets of monitoring data in the interval was calculated. The results are shown in Table 9.

[0113] Table 9 Concentration Time Distribution Series

[0114] Wind direction range Number of monitoring data groups included in this segment Average concentration data 0-30° 537 89 30-60° 467 118 60-90° 217 97 90-120° 179 256 120-150° 474 219 150-180° 575 236 180-210° 339 215 210-240° 153 149 240-270° 109 178 270-300° 267 119 <![CDATA[300 - 330°]]> 525 136 <![CDATA[330 - 360°]]> 478 115

[0115] Finally, the dispersion r of the average concentration data within each wind direction interval was calculated: the data are shown in Table 10.

[0116]

[0117] Among them, w v Let be the average concentration in the v-th wind direction interval, where v∈[1,q], and q is the number of wind direction intervals. To substitute all w into the calculation v The average value. The recommended number of concentration data points for the diagnostic monitoring data sequence is the number of real-time data points available from the device over 3-7 days.

[0118] Table 10 Average Concentration Across Wind Directions

[0119] <![CDATA[w1]]> <![CDATA[w2]]> <![CDATA[w3]]> <![CDATA[w4]]> <![CDATA[w5]]> <![CDATA[w6]]> <![CDATA[w7]]> <![CDATA[w8]]> <![CDATA[w9]]> <![CDATA[w 10 ]]> <![CDATA[w 11 ]]> <![CDATA[w 12 ]]> 89 118 97 256 219 236 215 149 178 119 136 115

[0120] Step 3: Complete the anomaly diagnosis based on the calculated sequences and whether there are pollution sources in the surrounding area.

[0121] like Figure 3 As shown, the specific method is as follows:

[0122] Step 1: If in the slope coefficient sequence, k i The percentage of data with a value ≥0 exceeds the threshold A or k. i If the percentage of data with a value ≤0 exceeds the threshold A, it is considered abnormal and the process ends; otherwise, proceed to step 2. The threshold A ranges from 80% to 100%.

[0123] Step 2: If in the trend coefficient sequence, g i The percentage of data with a value ≥0 exceeds the threshold B or g. i If the percentage of data with a value ≤0 exceeds the threshold B, it is considered abnormal and the process ends; otherwise, proceed to step 3. The threshold B ranges from 80% to 100%.

[0124] Step 3: Determine if there is a pollution source in the vicinity. If not, the instrument will function normally and the process will end. If so, proceed to Step 4.

[0125] Step 4: Determine the dispersion of the concentration time distribution sequence. If the obtained dispersion s is less than the threshold... If the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, proceed to step 5. The value of threshold C ranges from 5% to 50%.

[0126] Step 5: Determine the dispersion of the concentration-wind direction distribution sequence. If the obtained dispersion r is less than the threshold... If the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, the monitoring equipment is considered normal, and the diagnosis ends. The threshold D ranges from 5% to 50%.

[0127] Regarding the specific values ​​of A, B, C, and D, the following explanation uses the value of A as an example:

[0128] Install the instrument in the field environment, and under the premise of ensuring that the instrument is operating normally, run it for a long time and perform the following calculations a large number of times:

[0129] Construct a sequence of slope coefficients, and statistically analyze k for each slope coefficient sequence. i The percentage of data points ≥0 relative to the total number of data points in the sequence, and k i The percentage of data points ≤0 relative to the total number of data points in the sequence. For a large number of slope coefficient sequences, the maximum percentage is taken as A. For simplicity, it is recommended that A be between 80% and 100%.

[0130] Example 2 is for a monitoring device that can simultaneously monitor multiple substances:

[0131] Taking an air quality monitoring device as an example, the monitoring factors include six air parameters: SO2, NO2, O3, CO, PM10 and PM2.5. Since most of these devices are based on sensor principles, each monitoring factor uses different sensors to monitor the concentration of each factor in real time. Therefore, there may be a phenomenon where one sensor fails while the others are normal. Thus, it is necessary to diagnose each monitoring factor separately.

[0132] First, conduct equipment status parameter diagnosis, and then perform anomaly diagnosis on monitoring factors such as SO2, NO2, O3, CO, PM10 and PM2.5 respectively. If the diagnosis result of any monitoring factor is abnormal, the equipment operating status is abnormal.

[0133] Example 3 provides a simple method for identifying substances that are stable in the air and have relatively stable concentrations:

[0134] If the monitoring factor of a certain device is a gas that is commonly present in ambient air, a judgment method can be used when diagnosing the device. The specific method is described below.

[0135] Taking methane monitoring equipment as an example, it is known that methane is a common gaseous component in the air, and the concentration of methane in clean air under normal circumstances is about 1800-2200 ppb. If the monitoring object of a certain device is normal ambient air or polluted air around chemical enterprises, then theoretically the concentration of methane in the air should be greater than or equal to the normal content in clean air.

[0136] If the online monitoring data of a methane monitoring device shows that the methane concentration value has been consistently below 1600 ppb for a period of time, it indicates that the device is malfunctioning.

[0137] Example 4 addresses the scenario where real-time data updates from monitoring equipment are not timely due to communication problems:

[0138] Taking the online monitoring equipment for benzene series compounds at the plant boundary as an example, the measurement cycle for benzene series compounds is generally about 5-15 minutes. Here, we assume 5 minutes. In order to obtain the latest monitoring data from the equipment in a timely manner, the communication transmission interval is set to 5 minutes.

[0139] If, during normal operation, the device fails to acquire real-time data for 30 consecutive minutes (6 consecutive communication transmission cycles), the device is deemed to be malfunctioning. Specifically, this refers to a malfunction in the device system (device + power supply + communication links, etc.), such as a power outage causing the device to stop and be unable to transmit data, or a failure in the communication link between the device and the server.

[0140] Example 5: The equipment status can be determined based on the key status parameters of the monitoring equipment.

[0141] Taking the online monitoring equipment for benzene series compounds at the factory boundary as an example, most of them adopt the GC-FID detection principle. First, the sampling tube collects air samples, and then the low concentration of benzene series compounds in the air is enriched into high concentration of benzene series compounds through the cold trap. Finally, the benzene series compound concentration is measured by the GC-FID detector.

[0142] To improve the detection sensitivity of benzene series compounds, enrichment is an essential and important method under the condition of limited detector accuracy. For enrichment cold traps, temperature parameter is an important indicator and can be used for equipment operation status diagnosis.

[0143] For example, under normal circumstances, the temperature of the enriched cold trap is approximately (-25℃) to (-20℃). If the temperature of the enriched cold trap, as monitored in real time, is consistently higher than -10℃ during equipment operation, it indicates that the equipment is operating abnormally and requires further inspection and maintenance.

[0144] Example 6: Equipment Condition Diagnosis Method under Complex Environmental Conditions

[0145] Because gas concentration monitoring equipment is greatly affected by environmental factors, it is necessary to determine whether the instrument can operate normally in complex environments. For most environmental monitoring equipment, humidity is a significant factor affecting monitoring data. Here, we take humidity as an example to provide a method for judging the instrument's operating status in high humidity environments. If experimental conditions are available, humidity interference resistance experiments can be conducted directly in the laboratory, which will not be elaborated here. If the instrument is already installed on-site, big data analysis principles can be used to statistically analyze the changing patterns of monitoring data under different humidity conditions over a long period. If, under high humidity conditions, the concentration monitoring data remains consistently high (compared to monitoring data under normal environmental conditions) and remains so, the equipment is likely to be in an abnormal state under high humidity conditions.

[0146] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for dynamic diagnosis of abnormal states of a gas leak monitoring device, characterized in that, Includes the following steps: Step 1: Identify the type of gas being monitored by the gas leak monitoring equipment, obtain the concentration data, wind direction data, and time data of the monitoring equipment, and establish a monitoring data sequence to be diagnosed by combining the equipment detection limit and the normal concentration of the monitored gas in ambient air. Step 2: Based on the monitoring data sequence to be diagnosed, calculate the slope coefficient sequence of the concentration data, the trend coefficient sequence of the concentration data, the concentration time distribution sequence of the concentration data on the time data and its dispersion, and the concentration wind direction distribution sequence of the concentration data on the wind direction data and its dispersion. Step 3: Complete the anomaly diagnosis based on the calculated sequences and whether there are pollution sources in the surrounding area; The method for establishing the trend coefficient sequence is as follows: First, calculate the moving average of the concentration data from the j-th to the i-th data points in the monitoring data sequence to be diagnosed. When i > H, , When H≥i≥1, f0 = f1; H is the length of the moving window; The first in the monitoring data sequence to be diagnosed Concentration data; Then, calculate the trend coefficient sequence. : ， ； in, The moving average of the concentration data from the (j-1)th to the (i-1)th concentration data in the monitoring data sequence to be diagnosed. is the average concentration data in the monitoring data sequence to be diagnosed; M is the number of concentration data in the monitoring data sequence to be diagnosed. The method for step three is as follows: First, perform anomaly diagnosis on the slope coefficient sequence and trend coefficient sequence respectively. If either result is abnormal, the monitoring equipment is considered abnormal and the process ends. Otherwise, determine whether there is a pollution source around the monitoring equipment. If not, the monitoring equipment is considered normal and the process ends. If so, perform anomaly diagnosis on the concentration time distribution sequence and its dispersion, and the concentration wind direction distribution sequence and its dispersion respectively. If either result is abnormal, the monitoring equipment is considered abnormal and the process ends. Otherwise, the monitoring equipment is considered normal and the process ends.

2. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The method for constructing the monitoring data sequence to be diagnosed is as follows: The monitoring data sequence to be diagnosed includes concentration data, wind direction data, and time data when the monitoring data was acquired. If the normal concentration is greater than 3 times the equipment detection limit, then a diagnostic monitoring data sequence containing three-dimensional parameters including time data, wind direction data, and concentration data will be directly constructed. Otherwise, select concentration data with values ​​greater than the normal concentration, and combine them with wind direction data and time data at the same time as the concentration data to construct a monitoring data sequence to be diagnosed.

3. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The method for establishing the slope coefficient sequence is as follows: Slope coefficient sequence The calculation formula is as follows: ， ； in, The first in the monitoring data sequence to be diagnosed Individual concentration data, The first in the monitoring data sequence to be diagnosed Individual concentration data, ; is the average concentration data in the monitoring data sequence to be diagnosed; M is the number of concentration data in the monitoring data sequence to be diagnosed.

4. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The method for calculating the concentration time distribution sequence and its dispersion of concentration data over time is as follows: First, collect the concentration data of the monitoring device to be diagnosed over the past few days and establish a monitoring data sequence to be diagnosed; Then, the 0-24 hour period is divided into 24 or 12 time periods. Based on the time data in the monitoring data sequence to be diagnosed, the collected concentration data is divided into each time period, and the average value of multiple concentration data in each time period is calculated. Finally, the dispersion s of the average concentration data within each time period is calculated: ； in, The average concentration for the j-th time period is... p is the number of time periods. For all of the calculations The average value.

5. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The method for calculating the concentration-wind-direction distribution sequence of concentration data and its dispersion on wind-direction data is as follows: First, collect the concentration data of the monitoring device to be diagnosed over the past few days and establish a monitoring data sequence to be diagnosed; Then, the 0-360° wind direction is divided into 8, 12 or 16 wind direction intervals. Based on the wind direction data in the monitoring data sequence to be diagnosed, the collected concentration data is divided into each wind direction interval, and the average value of multiple sets of concentration data in each interval is calculated. Finally, the dispersion r of the average concentration data within each wind direction interval is calculated: ； in, Let be the average concentration in the v-th wind direction interval. q is the number of wind direction intervals. For all of the calculations The average value.

6. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The specific method for step three is as follows: Step 1: If in the slope coefficient sequence, The percentage of data exceeding threshold A or If the proportion of data exceeds the threshold A, it is judged as abnormal and the process ends; otherwise, step 2 is executed. Step 2: If in the trend coefficient sequence, The percentage of data exceeding threshold B or If the proportion of data exceeds the threshold B, it is judged as abnormal and the process ends; otherwise, step 3 is executed. Step 3: Determine if there is a pollution source in the vicinity. If not, the instrument will function normally and the process will end. If so, proceed to Step 4. Step 4: Determine the dispersion of the concentration time distribution sequence. If the obtained dispersion s is less than the threshold C* , If the concentration is the average value across all time periods, then the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, proceed to step 5. Step 5: Determine the dispersion of the concentration-wind direction distribution sequence. If the obtained dispersion r is less than the threshold D* , If the concentration is the average value across all wind directions, then the dispersion is insufficient, the monitoring equipment is malfunctioning, and the diagnosis ends; otherwise, the monitoring equipment is considered normal, and the diagnosis ends.

7. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 6, characterized in that, The threshold A ranges from 80% to 100%.

8. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 6, characterized in that, The threshold B ranges from 80% to 100%.

9. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 6, characterized in that, The threshold C ranges from 5% to 50%.

10. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 6, characterized in that, The threshold D ranges from 5% to 50%.

11. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 3, characterized in that, The number of M is not less than the number of real-time data obtained by the monitoring device to be diagnosed over 6 hours.

12. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, The length H of the moving window can be 3-6.

13. A method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 4 or 5, characterized in that, The number of concentration data in the monitoring data sequence to be diagnosed is the number of real-time data obtained by the monitoring device to be diagnosed over 3-7 days.

14. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, If the monitoring equipment can monitor multiple factors simultaneously, each factor needs to be diagnosed separately for abnormalities.

15. The method for dynamic diagnosis of abnormal states of a gas leak monitoring device according to claim 1, characterized in that, If the key status parameters of the monitoring equipment exceed the normal range, it is also considered abnormal. The key status parameters include the pressure, temperature, flow rate, current, and voltage of key components.

16. A dynamic diagnostic system for abnormal states of a gas leak monitoring device, employing the dynamic diagnostic method for abnormal states of a gas leak monitoring device as described in claim 1, characterized in that, It includes a monitoring data acquisition unit, a basic data acquisition unit, a data processing unit, and an anomaly diagnosis unit. The monitoring data acquisition unit and the basic data acquisition unit respectively input the acquired data into the data processing unit, and the data processing unit inputs the processed data into the anomaly diagnosis unit.

17. The abnormal state dynamic diagnosis system for a gas leak monitoring device according to claim 16, characterized in that, The monitoring data acquisition unit includes a concentration data acquisition module, a wind direction data acquisition module, and a time data acquisition module.

18. The abnormal state dynamic diagnosis system for a gas leak monitoring device according to claim 16, characterized in that, The basic data acquisition unit includes a device detection limit acquisition module and an ambient air normal concentration acquisition module.

19. The abnormal state dynamic diagnosis system for a gas leak monitoring device according to claim 16, characterized in that, The data processing unit includes a slope coefficient sequence processing module, a trend coefficient sequence processing module, a concentration time distribution sequence and its dispersion processing module, and a concentration wind direction distribution sequence and its dispersion processing module.

Images