An adaptive sampling method of ambient wind vector for near-field leakage tracing

By using an environmental wind vector adaptive sampling method, and leveraging the collaborative work of wind speed and direction sensors and concentration sensors, the problem of high accuracy and cost in reverse calculation caused by wind direction changes in traditional methods is solved, thus achieving efficient and low-cost leak tracing in complex atmospheric environments.

CN114490613BActive Publication Date: 2024-11-22IVOCMN SHANGHAI INTERNET OF THINGS TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202210113312.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-01-30
Publication Date
2024-11-22
Estimated Expiration
2042-01-30

AI Technical Summary

Technical Problem

Existing technologies cannot effectively adapt to complex atmospheric environments in near-field leak tracing, resulting in low accuracy, high cost, and low efficiency in back-calculation. In particular, when wind speed and direction are constantly changing, traditional methods cannot adjust the coordinate system in time, causing sensor nodes to be unable to accurately calculate the location of the leak source.

Method used

An adaptive sampling method for environmental wind vectors is adopted. Through the collaborative work of wind speed and direction sensors and concentration sensors, the sampling period is set according to the frequency of wind vector changes. Data cleaning and coordinate transformation are performed. Genetic algorithms and Nelder Mead simplex algorithms are used to optimize the inverse calculation process, ensuring that the sensor nodes in the model always meet the wind direction requirements, thereby improving the sensitivity and accuracy of data acquisition.

Benefits of technology

It improves the adaptability and accuracy of near-field leak tracing, reduces costs, solves the problem of insufficient sensor quantity, improves back-calculation efficiency, and ensures real-time back-calculation results in complex atmospheric environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114490613B_ABST
    Figure CN114490613B_ABST
Patent Text Reader

Abstract

The present invention discloses an environmental wind vector adaptive sampling method for near-field leakage tracing, comprising the following steps: 1. determining a reasonable sampling period; 2. collecting the wind speed U at the start time through a wind speed and direction sensor; n and wind direction α n ; 3. During the current sampling period, the concentration sensor collects the environmental concentration value C at multiple intervals n To C n+k ; 4. The wind speed and direction sensor is used to collect the wind speed U at the final moment of the sampling period n+1 and wind direction α n+1 ; 5. Environmental concentration value C n To C n+k Calculate the average value and according to the wind speed U n , wind direction α n , wind speed U n+1 Wind direction α n+1 6. Perform coordinate transformation; 7. Perform back calculation according to the average value of environmental concentration and wind vector coordinates; 8. Determine whether the traceability requirements are met. If yes, end the process; if no, return to step 2. The present invention better adapts to complex atmospheric environments, improves the accuracy of traceability back calculation, reduces traceability costs, and improves traceability efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of leakage tracing methods, and in particular to an environmental wind vector adaptive sampling method for near-field leakage tracing. Background Art

[0002] There are many dangerous chemicals in the chemical, petroleum, dye and some other industries, such as flammable and explosive toxic gases and liquids (such as liquefied petroleum gas, hydrogen sulfide, ammonia and chlorine, etc.). If these dangerous chemicals leak during the production or transportation process due to negligence or operational errors of the staff, it will inevitably cause great harm to the surrounding residents and the environment.

[0003] Dangerous chemical accidents are characterized by suddenness, which means they occur without any warning or preparation. When a dangerous chemical accident occurs, the harm to the human body and the environment often does not appear immediately, but is delayed and long-term, and the harm sometimes becomes greater over time; dangerous chemical accidents also have diverse forms and are very harmful, often causing heavy casualties and huge economic losses.

[0004] With the development of social economy, hazardous chemical explosion accidents have occurred frequently in recent years. The leakage and pollution of toxic chemicals have seriously threatened the safety of people’s lives and property, and even caused large-scale social panic and related social problems. Therefore, hazardous chemical explosion accidents have attracted widespread attention and concern from the society.

[0005] In order to reduce the serious harm and losses caused by dangerous chemical accidents, it is necessary to effectively control the pollution accidents caused by the leakage of toxic gases of dangerous chemicals. In addition to strengthening the prevention of sudden toxic environmental pollution accidents, it is also necessary to establish a sound emergency response system and improve the ability to deal with accidents. In recent decades, emergency personnel often use computer simulations based on mathematical models and the atmospheric diffusion process of leaked gases to describe the diffusion and transportation process of released substances in the environment, obtain the spatial and temporal distribution of the concentration of leaked substances, and provide a theoretical basis and basis for the relevant departments to carry out emergency evacuation of personnel and handle toxic substances. At the same time, on this basis, the scope of the hazard area with different degrees of severity can be determined, which is of great significance for the engineering design of dangerous chemicals and the formulation and implementation of emergency plans.

[0006] When toxic chemicals leak, obtaining relevant information about the leak source as early as possible is of great significance for controlling the development of the accident, formulating crowd evacuation strategies and emergency response. However, due to the characteristics of dangerous chemical accidents, it is difficult to obtain relevant information such as the source strength and location of the leak source. Therefore, inversion of the leak source and obtaining its key information has become a key issue in the chemical field.

[0007] The tracing method based on atmospheric model is currently a common rapid tracing method. First, a coordinate system is established in a scene where leakage may occur. The principle of establishment is that the wind direction is only along the x-axis. Assuming that the leakage source is located at the coordinate (x0, y0, z0) and the wind speed is constant, the concentration function C(Q) at the position (x, y, z) at time t can be obtained through the model. 0, x,x 0, y,y 0, z,z 0, u,t), Q0 is the leakage source strength; u is the wind speed. Using the model, establishing an inverse problem or directly based on probability statistics, using algorithms can quickly and real-timely trace the leakage source. In specific implementation, such as Figure 1 As shown, a sensor array is set up in the downwind area where the concentration may be captured, and the dominant wind direction is determined according to the meteorological data. The dominant direction is determined as the x-axis, and a coordinate axis is established. The sensor node coordinates in the coordinate axis, i.e. (x, y, z) and the concentration information c at the corresponding position are directly input into the inverse calculation program to obtain the source item information of the leakage source. The source item information includes the source intensity Q0 of the leakage source and the location of the leakage source (x0, y0, z0).

[0008] Another back-calculation method that is not based on atmospheric models is to use neural networks to train specific leakage scenarios. A model with strong generalization capabilities can be trained using massive amounts of leakage data, which can then be used for actual back-calculation. Since this back-calculation method is not based on mathematical and physical models, it is generally believed that this method has stronger adaptability to the environment.

[0009] Since the actual traceability system needs to be applied to actual scenarios, the wind speed or wind direction in actual scenarios is constantly changing, and the preset x-axis can only represent the main wind direction of the area. Especially in near-field traceability, the concentration field after the leak is more sensitive to changes in wind speed and wind direction. However, the traditional sampling method cannot adjust the coordinate system in time according to the change of wind vector, which results in the node coordinates collected at this time being only the coordinates in the preset coordinate system, rather than the coordinates under the influence of this wind. This violates the principle that the x-axis should be along the wind direction, resulting in the inverse calculation program being completely unable to accurately trace the source in the actual factory environment.

[0010] The neural network-based method has high requirements for data volume and data quality, and requires the collection of a large amount of real leakage data, which is not realistic. The training time in the early stage is too long, and each time a scenario is changed, it needs to be retrained, and the application cost is too high.

[0011] Therefore, the applicant has found a solution to the above-mentioned problem through beneficial exploration and research, and the technical solution to be introduced below is produced in this context. Summary of the invention

[0012] The technical problem to be solved by the present invention is: to provide an environmental wind vector adaptive sampling method for near-field leakage tracing that can better adapt to complex atmospheric environments, improve the accuracy of tracing back calculations, reduce tracing costs, and improve tracing efficiency in response to the shortcomings of the prior art.

[0013] The technical problem to be solved by the present invention can be achieved by adopting the following technical solutions:

[0014] An environmental wind vector adaptive sampling method for near-field leakage tracing comprises the following steps:

[0015] Step S10, determining a reasonable sampling period according to the wind vector change frequency at the leakage location;

[0016] Step S20: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the start time. n and wind direction α n ;

[0017] Step S30: In the current sampling period, the concentration sensor collects the environmental concentration value C at intervals for multiple times. n To C n+k ;

[0018] Step S40: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the final moment of the sampling period. n+1 and wind direction α n+1 ;

[0019] Step S50, the environmental concentration value C n To C n+k Calculate the average value to obtain the average value of the ambient concentration in the current sampling period. n , wind direction α n , wind speed U n+1 and wind direction α n+1 Perform coordinate transformation to obtain the wind vector coordinates of the current sampling period;

[0020] Step S60, performing back calculation processing according to the average value of the environmental concentration and the wind vector coordinates to obtain a leakage sampling data sample;

[0021] Step S70, determine whether the leakage sampling data sample meets the traceability requirements, if it is determined to be yes, then end, if it is determined to be no, then return to step S20.

[0022] In a preferred embodiment of the present invention, step S80 is further included between step S60 and step S70, in which the leakage sampling data samples obtained in step S60 are processed to narrow the search range so that the value range of the leakage sampling data samples obtained by processing is more accurate.

[0023] In a preferred embodiment of the present invention, in step S60, the back-calculation process based on the average value of the environmental concentration and the wind vector coordinates includes the following steps:

[0024] Step S61, constructing a back-calculation target function based on a diffusion model, and importing the average value of the ambient concentration and the wind vector coordinates into the back-calculation target function;

[0025] Step S62, using a genetic algorithm to perform a large-scale search in the solution space of the inverse objective function, to search for an initial point near the global optimal solution for the NelderMead simplex algorithm;

[0026] Step S63, using the Nelder Mead simplex algorithm to optimize the initial points near the searched global optimal solution, speeding up the optimization process, and finally finding the optimal solution, which is the leakage sampling data sample.

[0027] Due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0028] 1. The present invention uses sampling and data processing methods to disassemble the continuous release of smoke and rain into multiple releases of smoke. The information upload of the wind speed and direction sensor and the concentration sensor must establish a time axis. According to the frequency of wind speed and direction changes at the system installation location, a wind speed and direction sensor data upload cycle is set. Sufficient concentration information is uploaded within this cycle, and data cleaning is performed on these concentration information (usually the average method is used to reduce errors) to ensure that the real-time concentration situation is truly reflected. The data uploaded each time is the wind speed and the coordinates and corresponding concentrations of different nodes under the influence of the wind speed and wind direction. This is equivalent to treating continuous leakage as the superposition of multiple instantaneous leakages, and each sampling is equivalent to experiencing an instantaneous leakage.

[0029] 2. The present invention utilizes coordinate changes to make the preset coordinate system always meet the requirements of the atmospheric model. According to the time-varying wind direction, the direction of the x-axis is adjusted in time, so that the entire inverse calculation program always meets the principle of the model, that is, the x-axis is always along the wind direction, so that the position of the fixed sensor node in the model is more accurate, overcoming the weakness that the node cannot move with the wind direction, and improving the sensitivity of data collection to time-varying wind. It is different from the method of using robot nodes in other methods. This method has lower cost, simpler principle and more convenient execution.

[0030] 3. The present invention uses sampling and data processing to make the original back-calculation method better adapt to the complex atmospheric environment.

[0031] 4. The present invention can solve the problem of insufficient number of sensors, reduce costs and improve back-calculation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be 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.

[0033] Figure 1 It is a schematic diagram of the existing node arrangement and coordinate system construction based on the atmospheric model.

[0034] Figure 2 It is a schematic diagram of the process of the present invention.

[0035] Figure 3 It is a schematic diagram of the present invention applied to heavy gas release experiment B06.

[0036] Figure 4 It is a node effect diagram of heavy gas release experiment B06 after using the present invention.

[0037] Figure 5 It is a schematic diagram of the smoke puff trajectory at different times of the present invention.

[0038] Figure 6 2 is another schematic diagram showing the trajectory of smoke puffs at different times of the present invention.

[0039] Figure 7 It is a schematic diagram of the coordinate transformation of the present invention. DETAILED DESCRIPTION

[0040] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below with reference to specific diagrams.

[0041] See also Figure 2 , the figure shows an adaptive sampling method of ambient wind vector for near-field leakage tracing, including the following steps:

[0042] Step S10, determining a reasonable sampling period according to the wind vector change frequency at the leakage location;

[0043] Step S20: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the start time. n and wind direction α n ;

[0044] Step S30: In the current sampling period, the concentration sensor collects the environmental concentration value C at intervals for multiple times. n To C n+k ;

[0045] Step S40: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the final moment of the sampling period. n+1 and wind direction α n+1 ;

[0046] Step S50, the environmental concentration value C n To C n+k Calculate the average value to obtain the average value of the ambient concentration in the current sampling period. n , wind direction α n , wind speed U n+1 and wind direction α n+1 Perform coordinate transformation to obtain the wind vector coordinates of the current sampling period;

[0047] Step S60, performing back calculation processing according to the average value of the environmental concentration and the wind vector coordinates to obtain a leakage sampling data sample;

[0048] Step S70, determine whether the leakage sampling data sample meets the traceability requirements, if it is determined to be yes, then end, if it is determined to be no, then return to step S20.

[0049] A step S80 is also included between step S60 and step S70, in which the leakage sampling data samples obtained in step S60 are processed to narrow the search range, so that the value range of the leakage sampling data samples obtained by processing is more accurate.

[0050] In step S60, back calculation is performed based on the average value of the ambient concentration and the wind vector coordinates, including the following steps:

[0051] Step S61, constructing a back-calculation target function based on a diffusion model, and importing the average value of the ambient concentration and the wind vector coordinates into the back-calculation target function;

[0052] Step S62, global search phase: using a genetic algorithm to conduct a large-scale search in the solution space of the inverse objective function, to search for an initial point near the global optimal solution for the Nelder Mead simplex algorithm;

[0053] Step S63, enhanced search phase: using the Nelder Mead simplex algorithm to optimize the initial points near the searched global optimal solution, speeding up the optimization process, and finally finding the optimal solution, which is the leakage sampling data sample.

[0054] The purpose of the present invention is to improve the adaptability of the back-calculation system based on the atmospheric model to the time-varying wind vector from the perspective of sampling and data cleaning. At the same time, the present invention can also effectively solve the problem of insufficient sensor node distribution. Taking a heavy gas release experiment as an example, Figure 3As shown, in experiment B06, the optional monitoring points were equipped with concentration sensors corresponding to the released gas, a total of 12 nodes, and wind speed and direction sensors were placed at the environmental monitoring points. Figure 4 As shown, all of them are nodes that can be input into the back-calculation program for calculation, and it can be seen that the number of nodes has increased significantly. It can be seen that when the release takes a long time, the invention can solve the problem of insufficient number of sensors, reduce costs, and improve back-calculation efficiency.

[0055] When toxic and harmful gas components spread from the leakage source to the outside air, the main factors affecting their movement are the inertial force at the leak outlet, gravity and the transport effect of the atmosphere. Considering that the application scenario is a small leak of dangerous gas, the inertial force at the leak outlet can be excluded, and only the gravity and the transport effect of the atmosphere can be considered. For gases with a molecular weight less than that of air, that is, light gases, the atmospheric transport effect is dominant, while for gases with a molecular weight greater than that of air, that is, heavy gases, in their diffusion process, gravity is the dominant force at first, but as they are fully mixed and diluted with air, the gravity force is gradually replaced by the transport force of the atmosphere. Therefore, the transport effect of the atmosphere should also be considered during sampling. Since the inertial force is not considered, the component velocity of the dangerous gas on the x-axis can be considered to be the same as the wind speed at this moment, and the movement of the dangerous gas on the x-axis is synchronized with the change of the wind. Since the wind speed and wind direction are constantly changing and it is difficult to maintain a steady state, the present invention can regard the continuously leaked dangerous gas as the superposition of multiple instantaneous leaked smoke clouds, each of which is affected by different wind speeds and wind directions.

[0056] Since it is difficult to synchronize the concentration sensor with the wind speed and direction sensor, and the sampling frequencies are likely to be different, especially considering the transport effect of wind on the leaked gas cloud, when a puff of smoke just leaks out from the leak source, it has not yet affected the concentration sensor node that is farther away. It will affect the concentration measured by the node after a period of air transportation. Assume that at time t1, puff 1 starts to leak from the leak source, and the concentration sensor node A is d meters away from the leak source. The wind speed at this time is v1 meters / second. Then after d / v1 seconds, puff 1 is transported to node A by the wind. Assume that the next puff of smoke 2 leaks from the leak source after △t seconds. The wind speed at this time is v2 meters / second. Figure 5 As shown, during the time period from t1+d / v1 to t1+△t+d / v2, it can be said that node A has been affected by smoke puff 1. The concentration measured in each response at this time is the concentration of smoke puff 1. △t at this time is the sampling period of the wind speed and direction sensor.

[0057] In this cycle, the average value of the concentration sensor is taken as the concentration value in this cycle, thereby reducing the measurement error of the concentration sensor to a certain extent. One of the conditions for this sampling method to be valid is d / v1<△t+d / v2, that is, the next puff will not affect the node earlier than the first puff, and the possibility of overlapping of the previous and next puffs should be avoided as much as possible. In actual experiments, it is easy to find that after each sampling cycle of the wind speed and wind direction sensor, it is not only the wind speed, such as Figure 7 As shown, the wind direction also changed significantly, which also meant that there was less overlap between the two smoke puffs.

[0058] In order to make this method further adapt to the situation of constantly changing wind direction in the natural environment, each time the wind speed and wind direction sensor responds, the coordinate system must be changed according to the responded wind direction so that the x-axis always points to the wind direction, and finally all the coordinates at different times and the corresponding concentrations are summarized into the same coordinate system. In this way, there is no need to worry about whether the wind direction at a certain moment deviates too much from the x-axis of the design coordinate system, resulting in a large error, and more and more accurate sample data can be obtained. The present invention brings the natural non-ideal state closer to the ideal state in time and space, which is more conducive to modeling and back calculation.

[0059] The present invention uses sampling and data processing methods to disassemble the continuously released smoke plume into multiple released smoke puffs. The information upload of the wind speed and direction sensor and the concentration sensor must establish a time axis. A wind speed and direction sensor data upload cycle is set according to the frequency of wind speed and direction changes at the system installation location. Sufficient concentration information is uploaded during this cycle, and data cleaning is performed on these concentration information (usually by averaging to reduce errors) to ensure a more realistic reflection of the real-time concentration situation. The data uploaded each time is the wind speed and the coordinates and corresponding concentrations of different nodes under the influence of the wind speed and wind direction. This is equivalent to treating continuous leakage as the superposition of multiple instantaneous leakages, and each sampling is equivalent to experiencing an instantaneous leakage.

[0060] The present invention utilizes coordinate changes to make the preset coordinate system always meet the requirements of the atmospheric model. According to the time-varying wind direction, the direction of the x-axis is adjusted in time, so that the entire back-calculation program always meets the principle of the model, that is, the x-axis must be along the wind direction, so that the position of the fixed sensor node in the model is more accurate, overcoming the weakness that the node cannot move with the wind direction, and improving the sensitivity of data collection to time-varying wind. It is different from the method of using robot nodes in other methods. This method has lower cost, simpler principle and more convenient execution.

[0061] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An environmental wind vector adaptive sampling method for near-field leakage tracing, characterized in that: The following steps are involved: Step S10, determining a sampling period according to the wind vector change frequency at the leakage location; Step S20: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the start time. n Wind direction α n ; Step S30: In the current sampling period, the concentration sensor collects the environmental concentration value C at intervals for multiple times. n To C n+k ; Step S40: In the current sampling period, the wind speed and direction sensor is used to collect the wind speed U at the final moment of the sampling period. n+1 and wind direction α n+1 ; Step S50, the environmental concentration value C n To C n+k Calculate the average value to obtain the average value of the ambient concentration in the current sampling period. n , wind direction α n , wind speed U n+1 and wind direction α n+1 Perform coordinate transformation to obtain the wind vector coordinates of the current sampling period; Step S60, performing back calculation processing according to the average value of the environmental concentration and the wind vector coordinates to obtain a leakage sampling data sample; Step S70, judging whether the leakage sampling data sample meets the traceability requirements, if it is judged to be yes, then the process ends, if it is judged to be no, then the process returns to step S20; Also included between step S60 and step S70 is step S80, in which the leakage sampling data samples obtained in step S60 are processed to narrow the search range, so that the value range of the leakage sampling data samples obtained by the processing is more accurate; In step S60, the back calculation process is performed based on the average value of the environmental concentration and the wind vector coordinates, including the following steps: Step S61, constructing a back-calculation target function based on a diffusion model, and importing the average value of the ambient concentration and the wind vector coordinates into the back-calculation target function; Step S62, using a genetic algorithm to perform a large-scale search in the solution space of the inverse objective function, to search for an initial point near the global optimal solution for the Nelder Mead simplex algorithm; Step S63, using the Nelder Mead simplex algorithm to optimize the initial points near the searched global optimal solution, speeding up the optimization process, and finally finding the optimal solution, which is the leakage sampling data sample.

Citation Information

Patent Citations

  • Dangerous chemical gas leakage isoconcentration line mass center track tracing algorithm

    CN108920818A

  • Concurrent in-situ measurement of wind speed and trace gases on mobile platforms for localization and qualification of emissions

    US20210190745A1