Dynamic environment video AI analysis early warning system for simulating neural network calculation model

Abstract

The invention discloses a dynamic environment video AI analysis early warning system for simulating a neural network calculation model, and the system employs a machine learning algorithm to build an error state recognition model, a pipeline leakage recognition model, and employs static target data, dynamic target data, logic events and behavior target data analysis data as training samples. Multi-algorithm fusion is researched to improve the dynamic environment monitoring accuracy and the pipeline leakage monitoring precision, so that an error state recognition model and a pipeline leakage recognition model are continuously optimized through machine learning; by means of the optimized and improved model, dynamic environment data such as personnel violation behaviors, equipment operation states, animal behaviors and pipeline valve leakage are processed and analyzed, and the management process of accurate alarm, graded pushing and closed-loop disposal is established. The system ensures that violation behaviors, equipment with abnormal operation states and oil and gas leakage events are found in time and early warned in time, achieves the purposes of rapid control and rapid disposal, and effectively reduces production accidents and safety and environmental protection risks.

Description

Technical Field

[0001] The present invention relates to the technical field of oilfield dynamic environment early warning, and more specifically, to a dynamic environment video AI analysis and early warning system that simulates a neural network computing model. Background Art

[0002] With the rapid development of the national economy and the continuous acceleration of urbanization, my country's demand for energy is increasing. As the pillar of energy, the oil and gas industry is related to the country's economic lifeline. It is of great significance to carry out relevant technical research on oil and gas exploration and development.

[0003] In recent years, with the continuous expansion of the scope of oil and gas development, the number of oil and gas field gathering and transportation pipelines is huge, the distribution area is wide, the terrain elevation difference is large, the medium is oil, water and gas multiphase mixed, the pipe is not fully filled, and the oil is not continuously transported. The working conditions are complex, especially for the booster station and oil transfer station gathering and transportation pipelines below the joint station (less than ), it is necessary to continuously monitor various equipment conditions, personnel behaviors, leakage conditions, etc., and the above-mentioned monitoring objects are mostly dynamic. The existing technology mainly collects various types of data by installing various sensors and cameras, and then conducts routine analysis on the data to obtain corresponding processing instructions. Research has found that when processing static data, the existing technology can obtain more accurate processing results and give correct processing instructions, but when processing dynamic data and videos, misjudgments often occur, and it is impossible to accurately warn of violations, equipment with abnormal operating conditions, and oil and gas leaks. Summary of the invention

[0004] The purpose of the present invention is to provide a dynamic environment video AI analysis and early warning system that simulates a neural network computing model, so as to solve the problem that the prior art is unable to accurately process dynamic data and videos in the working environment of oil and gas fields and thus cannot accurately warn of violations, equipment with abnormal operating conditions, and oil and gas leaks.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a dynamic environment video AI analysis and early warning system that simulates a neural network computing model, comprising a collection line front-end data acquisition module, a preprocessing module, an operation and analysis module, an early warning module and a monitoring module, the collection line front-end data acquisition module is connected and coordinated with the preprocessing module, the preprocessing module is connected and coordinated with the operation and analysis module, the operation and analysis module is connected and coordinated with the early warning module, and the monitoring module is connected and coordinated with the collection line front-end data acquisition module.

[0006] The overall information of the cluster pipeline is collected through the front-end data acquisition module of the pipeline collection line, and the data in all directions is processed and analyzed using the preprocessing module. The pattern classification of the neural network is established, and then the various data are calculated through the operation and analysis module. When an abnormality occurs, the early warning module pushes an alarm.

[0007] Preferably, the data acquisition module at the front end of the collection line includes a sensor unit, a pattern recognition unit, a communication unit and a dynamic environment video recognition unit.

[0008] Preferably, the operation and analysis module includes a fuzzy logic unit, a neural network unit, an expert system connection unit, a rough set theory unit and an artificial intelligence unit.

[0009] Preferably, the monitoring module includes a comprehensive flow balance unit, a negative pressure wave monitoring unit and a distributed optical fiber monitoring unit.

[0010] Preferably, the operation and analysis module includes a machine learning algorithm to establish a pipeline leakage identification model and an intelligent flow balance algorithm.

[0011] Preferably, the preprocessing module is used to process and analyze information such as pipeline flow, pressure, pump frequency, etc.

[0012] Preferably, the early warning module establishes a management process of accurate alarm, graded push, and closed-loop disposal.

[0013] Preferably, the dynamic environment video recognition unit adopts multi-scale target detection, and cooperates with the pattern recognition unit to respectively perform static target recognition, dynamic target recognition, logical event recognition and behavioral target recognition.

[0014] Technical effects and advantages of the present invention:

[0015] The present invention provides a dynamic environment video AI analysis and early warning system that simulates a neural network calculation model. A machine learning algorithm is used to establish an error state recognition model and a pipeline leakage recognition model. Static target data, dynamic target data, logical events and behavioral target data analysis data are used as training samples. Multi-algorithm fusion is studied to improve the accuracy of dynamic environment monitoring and the accuracy of pipeline leakage monitoring, so that the error state recognition model and the pipeline leakage recognition model are continuously optimized through machine learning. With the help of the optimized and improved model, dynamic environment data such as personnel violations, equipment operating status, animal behavior, pipeline valve leakage, etc. are processed and analyzed. By establishing a management process of accurate alarm, hierarchical push, and closed-loop disposal, it is ensured that violations, equipment with abnormal operating status, and oil and gas leakage incidents are discovered in time and warned in time, so as to achieve the purpose of rapid control and rapid disposal, and effectively reduce production accidents and safety and environmental protection risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall process structure of the present invention.

[0017] Figure 2 It is a schematic diagram of the data acquisition module at the front end of the collection line of the present invention.

[0018] Figure 3 This is a comprehensive identification flow chart of the leakage identification model of the present invention.

[0019] Figure 4 It is a schematic diagram of the working process of the error state recognition model of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Example 1

[0022] A dynamic environment video AI analysis and early warning system simulating a neural network computing model, the structure is as follows Figure 1-4 As shown, it includes a front-end data acquisition module for the collection line, a preprocessing module, an operation and analysis module, an early warning module and a monitoring module. The front-end data acquisition module for the collection line is connected and coordinated with the preprocessing module, the preprocessing module is connected and coordinated with the operation and analysis module, the operation and analysis module is connected and coordinated with the early warning module, the monitoring module is connected and coordinated with the front-end data acquisition module for the collection line, and the operation and analysis module includes a machine learning algorithm to establish a pipeline leakage identification model and an intelligent transmission balance algorithm.

[0023] The overall information of the cluster pipeline is collected through the front-end data acquisition module of the pipeline collection line, and the data in all directions is processed and analyzed using the preprocessing module. The pattern classification of the neural network is established, and then the various data are calculated through the operation and analysis module. When an abnormality occurs, the early warning module pushes an alarm.

[0024] Example 2

[0025] The data acquisition module at the front end of the collection pipeline includes a sensor unit, a pattern recognition unit, a communication unit and a dynamic environment video recognition unit. The dynamic environment video recognition unit adopts multi-scale target detection and cooperates with the pattern recognition unit to respectively recognize static targets, dynamic targets, logical events and behavioral targets.

[0026] In specific implementation, the dynamic environment video recognition unit is used to shoot static targets, dynamic targets, logical events and behavioral targets respectively, and the transmission unit is used to transmit data. The classified static targets are identified through the pattern recognition unit, the work clothes worn by personnel are identified, the model and license plate of the vehicle are identified, and various areas such as warning areas, danger areas and restricted areas are identified and monitored, and the pumping units, pipelines, booms, escape signs, etc. in the equipment are identified;

[0027] Identify dynamic targets, identify the protective gear worn by personnel, identify and record the location, number and retention time of vehicles, identify and record the number, location and retention time of animals, identify wellhead leakage, ground leakage, pipeline leakage, valve dripping, etc., identify pump stoppage, oil and water tank overflow, etc., and identify smoke and fire occurrences;

[0028] Identify behavioral targets, such as people smoking, making phone calls, and gatherings; identify vehicles wandering around, whether they have drivers, and illegal parking; identify equipment pumping units turning forward and backward, hot work equipment falling to the ground, and identify operations such as throwing balls, cleaning, and adding chemicals during operations;

[0029] Identify logical events, whether personnel are on patrol, off duty, sleeping on duty, etc., identify whether the driver of the vehicle has parked illegally, identify the rationality of the escape signs in the equipment and the road crossing protection, identify water dripping from the valve room and the value of the valve room instrument returning to zero, identify whether the spacing between hot work equipment is insufficient during the operation, the fire-fighting equipment is not detected, the safety rope is not tied during high-altitude operations, people are not standing under the hanging objects, the crane support legs are not extended, the excavation operation soil pile is too high, and the pit is too shallow.

[0030] Example 3

[0031] The operation and analysis module includes a fuzzy logic unit, a neural network unit, an expert system connection unit, a rough set theory unit and an artificial intelligence unit.

[0032] The neural network pattern classification is established through the neural network unit, and the fuzzy logic unit uses the rough set theory to simplify the fuzzy rules, and the key knowledge base such as fault characteristics is extracted from the expert system connection unit.

[0033] Example 4

[0034] The pre-processing module is used to process and analyze information such as pipeline flow, pressure, pump frequency, etc., and the early warning module establishes a management process of accurate alarm, graded push, and closed-loop disposal.

[0035] Machine learning algorithms are used to establish error state recognition models and pipeline leakage recognition models. Static target data, dynamic target data, logical events, and behavioral target data analysis data are used as training samples. Multi-algorithm fusion is studied to improve the accuracy of dynamic environment monitoring and pipeline leakage monitoring, so that the error state recognition model and pipeline leakage recognition model are continuously optimized through machine learning. With the help of the optimized and improved model, personnel violations, equipment operating status, animal behavior, pipeline valve leakage, etc. are processed and analyzed. By establishing a management process of accurate alarm, hierarchical push, and closed-loop disposal, it is ensured that violations, equipment with abnormal operating status, and oil and gas leakage incidents are discovered in time and warned in time, so as to achieve the purpose of rapid control and rapid disposal, effectively reducing production accidents and safety and environmental risks.

[0036] 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 spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic environment video AI analysis and early warning system that simulates a neural network computing model, Features: It includes a front-end data acquisition module for the collection line, a preprocessing module, an operation and analysis module, an early warning module and a monitoring module. The front-end data acquisition module for the collection line is connected and coordinated with the preprocessing module, the preprocessing module is connected and coordinated with the operation and analysis module, the operation and analysis module is connected and coordinated with the early warning module, and the monitoring module is connected and coordinated with the front-end data acquisition module for the collection line.

2. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The data acquisition module at the front end of the collection pipeline includes a sensor unit, a pattern recognition unit, a communication unit and a dynamic environment video recognition unit.

3. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The operation and analysis module includes a fuzzy logic unit, a neural network unit, an expert system connection unit, a rough set theory unit and an artificial intelligence unit.

4. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The monitoring module includes a comprehensive flow balance unit, a negative pressure wave monitoring unit and a distributed optical fiber monitoring unit.

5. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The operation and analysis module includes a machine learning algorithm to establish a pipeline leakage identification model and an intelligent transmission balance algorithm.

6. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The preprocessing module is used to process and analyze information such as pipeline flow, pressure, pump frequency, etc.

7. According to claim 1, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The early warning module establishes a management process of accurate alarm, graded push and closed-loop disposal.

8. According to claim 2, a dynamic environment video AI analysis and early warning system simulating a neural network computing model, Features: The dynamic environment video recognition unit adopts multi-scale target detection and cooperates with the pattern recognition unit to respectively perform static target recognition, dynamic target recognition, logical event recognition and behavioral target recognition.

Images