In recent years, with the rapid development of industrial Internet, big data, artificial intelligence and other emerging technologies, unmanned aerial vehicle (UAV) with its high flexibility and low cost, has a broad application demand in the field of inspection and detection in dangerous areas or space that is inconvenient for human to reach. At present, the application of UAV in outdoor scene flight is relatively mature, which can be deeply applied in agriculture, forestry and plant protection, power line patrol, oil pipeline inspection, land surveying and mapping, artificial rainfall, aviation remote sensing, emergency rescue, geological survey, security patrol, logistics express, medical rescue, marine remote sensing, news reporting and many other industrial scenes. However, in the complex enclosed space, especially in the high-altitude, high dust, double-blind (no GPS, no light) industrial scene similar to the boiler of thermal power plant, there are many technical problems in intelligent flight, high-definition shooting and image processing of UAV, which has become one of the most important issues to be solved in the industrial UAV industry. The following will be highly representative of the thermal power plant boiler furnace inspection UAV as an example, related to the elaboration and research.
Application requirements of intelligent inspection UAV for boiler in 1 Thermal Power Plant
After more than 20 years of leaping development of thermal power generating units, China has developed into the country with the largest installed capacity in the world. In recent years, how to integrate the new technological revolution of big data and artificial intelligence, improve the level of equipment management and safety production, and promote the intelligent transformation and industrial upgrading of thermal power plants, has become an important issue to be solved urgently in the thermal power industry.
The boiler of thermal power plant is one of the core power generation equipment. Its volume is huge, especially the ultra supercritical million class boiler, its maximum is about 100m, the internal span can reach 60m, the depth is about 40m. The environment of pulverized coal combustion in the boiler is complex and harsh, which requires long-term stable operation under the condition of ultra supercritical high temperature and high pressure parameters of steam. Therefore, the requirements of equipment reliability and inspection quality are very high, so it is of great significance to ensure the safe operation of the boiler. For boiler shutdown detection, manual detection method is generally adopted after setting up large scaffold or lifting platform. The inspector found the corrosion, crack, deformation, wear and other defects in the boiler through visual inspection or instrument on the scaffold. This traditional detection method needs to build a large number of scaffolds and other preparatory work, with huge workload, long construction period, high risk coefficient, and high requirements for the skills and experience of inspectors. Therefore, the development and design of industrial UAV suitable for this confined space, using it to carry out autonomous flight and high-definition photography of the boiler furnace interior, transmit the image to the outside in real time for calculation and processing, intelligently identify internal defects and other deterioration problems and conduct real-time evaluation. This upgrading and transformation has great attraction for traditional thermal power plants.
Research on UAV technology at home and abroad
2.1 application status of UAV
Commercial UAVs are generally divided into two categories: one is consumer oriented UAV for ordinary consumers, which has a large market scope, low manufacturing cost and low technical threshold. Chinese UAV manufacturers occupy more than 70% of the global market share in this field; The second category is industrial UAV, mainly for specific industry users, which needs to have high technology, quality and reliability. Therefore, the R & D cycle is long and the investment is high.
2.2 research status of UAV navigation and perception technology
The flight of UAV needs navigation, that is to correctly guide the UAV to arrive at the destination in the specified time along the predetermined track according to the specified accuracy requirements. In order to make the UAV successfully complete the scheduled navigation task, it is necessary to know the real-time position, navigation speed, heading and other navigation parameters of the UAV. At present, the navigation technologies used in UAV mainly include inertial navigation, satellite navigation, Wi Fi indoor navigation, Bluetooth indoor positioning technology, ultra wideband UWB indoor positioning technology, visual navigation, lidar navigation, etc.
2.3 Application Status of UAV in confined space
At present, the international community has initially carried out the research of UAV inspection work in the closed industrial scene, mainly through the form of adding protective cover outside the fuselage, which can effectively protect the safe flight of UAV. However, in order to make UAV flight inspection more efficient in boiler furnace and other confined space, the following problems need to be overcome and solved.
1) The UAV cannot hover in the boiler. At present, UAVs usually use satellite navigation or visual navigation. However, in the “double-blind” environment of boiler satellite signal rejection and darkness, satellite navigation and visual navigation can not be used, and unmanned aerial vehicle (UAV) cannot carry out autonomous positioning. Therefore, the two UAVs in the boiler chamber detection operation, put forward higher requirements for operators. Especially when the UAV is flying in the boiler furnace, the operator can not observe the UAV, so it is difficult to operate.
2) Unmanned aerial vehicle (UAV) is unable to realize three-dimensional autonomous obstacle avoidance in boiler. At present, the two UAVs can avoid obstacles in the form of protective cover, which has certain shortcomings: on the one hand, some slender and protruding structures in the boiler may pass through the protective cover and affect the flight safety; on the other hand, when the protective cover collides with the furnace wall, although it will not cause flying accidents, it will affect the operation experience and image quality. Therefore, it is necessary to study the active obstacle avoidance method based on sensor and intelligent algorithm.
3) Unable to determine the location of the UAV image in the boiler. At present, it is difficult for the two UAVs to correspond to the boiler body position after shooting the boiler images. Due to the high similarity of different parts in the furnace, even if defects are found, the specific location of defects can not be known after recording a large number of images. Therefore, it is necessary to study the relative positioning method of UAV in the boiler, on this basis, calculate the corresponding position of the image in the furnace body, so as to form an effective record of the image.
4) The aircraft has a short endurance. At present, the flight time of UAV is in
It can not meet the requirements of long-distance and long-time flight between 10min and 15min. Once the power is lost, the battery should be replaced frequently. For the boiler with high furnace, the flight time should be estimated during flight. If the flight is interrupted and the battery needs to be replaced, it is difficult to complete the detection task.
3. Research and development objectives of intelligent inspection UAV in confined space
When the UAV is flying in the boiler furnace, the UAV is in the “double blind” environment of satellite rejection and no visible light. This environment puts forward high technical innovation requirements for the UAV’s perception, navigation and control technology. The research and development objectives are as follows:
3.1 precise navigation of UAV in complex environment without satellite signal, illumination and strong electromagnetic interference
The boiler furnace belongs to the metal enclosed environment, and the UAV is in the “double blind” environment without satellite signal and light when carrying out boiler detection. At present, the realization of intelligent flight control, stabilized hovering, route planning, obstacle avoidance and other functions of commercial UAV are based on the premise of wide vision or good light. Through GPS, inertial navigation and artificial vision module, UAV can obtain position, attitude, speed, obstacle distance and other data for comprehensive calculation, so as to realize intelligent functions such as stability enhancement, hovering, obstacle avoidance and route planning. But in the boiler, the GPS signal is blocked, the compass is disturbed by metal structure, and the vision sensor cannot work because of the dark light. Therefore, in order to ensure the safe flight of the aircraft, it is necessary to build an autonomous navigation method for unmanned aerial vehicle (UAV) based on the new navigation technology.
3.2 obstacle perception and autonomous obstacle avoidance in non light environment of boiler
The boiler furnace inspection is mainly a close-up inspection task, which means that through close observation of the boiler surface, the problems such as weld cracking and surface defects are investigated. In order to obtain high-resolution image information of furnace body, the aircraft needs to fly close to the furnace wall. In order to ensure the safe flight of aircraft, it is necessary to accurately perceive and model the complex structure in the furnace. At the same time, the UAV should be able to sense the environment independently, plan the path reasonably and avoid obstacles in time, otherwise there is the risk of collision between UAV and furnace wall, which will lead to accident.
3.3 UAV system design for complex boiler detection task
In order to detect the cracks and cracks in the furnace, the unmanned aerial vehicle (UAV) is required to detect the cracks and cracks in the furnace, and to detect the cracks and cracks in the furnace. However, during the inspection process, it is easy to be affected by light, floating dust, electromagnetic interference and other adverse factors, so it is difficult for UAV to hover stably and take clear and reliable detection images; in addition, the detection task is heavy and time-consuming, and the battery power supply system used by UAV is difficult to meet the requirements of UAV for work efficiency. Therefore, it is necessary to design the hardware system structure of the UAV for boiler inspection based on the requirements of UAV perception, navigation and control.
3.4 Autonomous Fault Tolerant Navigation of UAV in complex boiler environment
The intelligent flight control, stabilized hovering, route planning, obstacle avoidance and other functions of traditional commercial UAV are based on the premise of open or good light. In principle, the UAV obtains the current position information, altitude information, attitude data, current speed and obstacle distance data through GPS, inertial navigation and artificial vision module for comprehensive calculation, so as to realize intelligent functions such as stability enhancement, hovering, obstacle avoidance and route planning. However, in the furnace, the GPS signal is blocked, the compass is disturbed by metal structure, and the artificial vision module cannot work because of the dim light, which leads to the loss of some intelligent functions of the UAV, and there will be obvious drift in the flight, so the ground controllers need to rely on rich experience to control the flight attitude. In the dark and closed environment, UAV control difficulty and operation safety risk increase. The reliable positioning of UAV in this environment has become a prominent problem, and it is one of the difficulties to solve the autonomous positioning in this environment.
3.5 boiler detection simulation and data reproduction based on 3D model
In the process of UAV development, it is necessary to carry out a large number of flight control system semi physical real-time simulation and scientific research flight test, and the observation, analysis and processing of flight data is a very important link. However, with the increase of flight data, this simple method can not meet the new requirements. At the same time, the boiler field detection usually needs a lot of energy, and the real data of the sensor is often hard to come by. The construction of boiler detection simulation environment and the simulation and reproduction of boiler data can reduce the workload.
4. Key technologies for inspection and development of intelligent UAV in confined space
The boiler furnace environment is a “double blind environment” without GPS and illumination. The traditional positioning method based on GPS or visual sensor is not suitable for this environment. Therefore, UAV high-precision positioning is one of the key technologies of this project. All boilers are made of metal, which will have a great impact on the communication between UAV and ground terminal. In addition, the flight time of traditional UAV is affected by the battery, and it often needs to fly repeatedly to realize the inspection of the whole boiler. Therefore, the communication and power supply of UAV in furnace is also one of the key technologies of this project.
To sum up, the key technologies of UAV for boiler furnace autonomous flight detection of thermal power plant are as follows:
4.1 UAV positioning technology based on 3D lidar
In the closed boiler environment with extremely poor illumination conditions, the traditional semantic composition technology based on visual sensor is not suitable for this environment. As an autonomous sensing sensor, lidar is not affected by the surrounding lighting conditions, with high ranging accuracy and wide measurement range. Therefore, this project uses lidar as the main sensing means of UAV. Through the research of lidar real-time composition technology, the error principle is analyzed, and the accurate three-dimensional map of boiler is constructed. The research contents include: probability expression of raster map based on LIDAR point cloud, and map model construction technology based on 3D point cloud information.
4.2 furnace obstacle perception and identification technology based on 3D lidar
3D lidar can get real-time obstacle information around UAV by laser ranging principle. The collected lidar information is not only used for UAV navigation and positioning, but also for intelligent perception of boiler environment. Therefore, this project is based on the real-time radar obstacle avoidance function of the unmanned aerial vehicle (UAV). The research contents include: Lidar Point Cloud perception and processing, furnace obstacle self identification technology.
4.3 asynchronous heterogeneous sensor multi-source information fusion technology
In order to realize the stable and safe flight of UAV in the complex environment of boiler, it is impossible to rely on a single sensor. From the positioning point of view, a single sensor can not meet the high-frequency and high-precision pose calculation required by UAV control; from the perspective of safe flight, single sensor information is limited, which can not meet the omni-directional detection of UAV surrounding obstacles. Based on the above analysis, it is necessary to study a variety of asynchronous heterogeneous sensor information fusion algorithms to achieve high-precision pose calculation and UAV environment perception.
4.4 UAV radar point cloud data and boiler model matching technology
UAV can sense the distance information in the furnace through 3D lidar, thus forming point cloud data. Through the correlation matching algorithm, the point cloud data is matched with the furnace model in the boiler, and the relative position of UAV in the boiler can be obtained. Furthermore, through the distance and angle information between the camera and the furnace, the corresponding position of the image in the boiler can be calculated, so as to provide accurate reference for detection.
4.5 power supply scheme design of UAV in confined space
Boiler inspection is a long-term operation task, and UAV usually needs to stay in the air for an hour or even longer in a flight operation. For this special long-time operation task, it is necessary to design the power supply scheme of UAV. The research contents include: UAV cable power supply system design based on tethered UAV solution, UAV airborne standby power supply scheme design considering ground to air power transmission fault, and ground power supply design of high-power and high reliability UAV system.
The use of intelligent UAV inspection in confined space boiler furnace has significant advantages of low cost, high efficiency and high intelligence, which plays a great role in the efficient detection of boiler furnace, which has great value to enhance the intelligence of traditional thermal power generation industry.
With the development of technology, the UAV platform will be equipped with thickness measurement equipment, spectral analysis equipment, etc., which can detect the boiler more comprehensively, so as to achieve the purpose of replacing manual work. In addition, the UAV related technologies obtained by this project can also be applied in the industrial environment detection of cabin, nuclear power station, oil tank, large pipeline and other confined spaces.
Editor in charge: GT