A digital twin driven new energy power station inspection method

Abstract

This invention discloses a digital twin-driven inspection method for new energy power plants, comprising the following steps: selecting and installing various types of sensors at key locations of critical equipment in the power plant; selecting suitable camera placement locations for different locations within the power plant using algorithms based on coverage and resolution optimization; completing a digital twin power plant performance model using a combination of laser scanning and camera imaging; forming an intelligent data display interface based on rule-based reasoning and machine learning, including fault prediction, diagnosis, and alarm mechanisms, thereby building a digital twin system; uploading information collected by sensors and cameras to the digital twin system using appropriate communication technologies; and completing the power plant inspection by using an inspection robot and cameras to collect power plant data. This invention, without increasing the cost budget, can simultaneously meet the requirements of real-time performance and accuracy, providing strong support for the safe and stable operation of new energy power plants.

Description

Technical Field

[0001] This invention relates to the field of new energy power plant inspection technology, specifically to a digital twin-driven new energy power plant inspection method. Background Technology

[0002] Photovoltaic power generation, with its low cost and convenient installation, has become an important sector in the new energy field and has experienced rapid development. However, in order to maximize the use of sunlight, photovoltaic power plants are often built in vast areas such as mountains and deserts, or installed in special locations such as the surface of water or the rooftops of buildings. These unique geographical environments bring enormous challenges to the safety inspection and maintenance of photovoltaic power plant equipment.

[0003] Traditional methods for detecting equipment safety faults in photovoltaic power plants involve dispatching personnel to the site for inspection only when a fault becomes severe enough to trigger a system alarm. This approach is not only time-consuming and has significant time lag, but it is also inefficient and fails to meet the demands of modern photovoltaic power plants for high-efficiency operation. With the rapid development of information technology, artificial intelligence algorithms such as image processing and deep learning have been introduced into this field. However, due to the concealed nature of photovoltaic inverter faults, ordinary camera technology cannot directly capture the characteristics of such faults, leading to a decrease in detection accuracy. Furthermore, these algorithms have significant drawbacks: on the one hand, they require upgrading existing hardware, increasing additional investment costs; on the other hand, their real-time performance is unsatisfactory, failing to detect and address faults promptly. Summary of the Invention

[0004] Purpose of the invention: In order to overcome the shortcomings of the prior art, the present invention provides a digital twin-driven method for inspecting new energy power plants.

[0005] Technical solution: The digital twin-driven inspection method for new energy power plants provided by this invention includes the following steps: S1. Select and install various types of sensors at key locations of critical equipment in the power plant; S2. For different locations in the power station, use algorithms based on coverage and resolution optimization to plan and select suitable camera deployment locations. S3. By combining laser scanning and camera imaging, a digital twin power plant performance model is completed; rule-based reasoning and machine learning are used to form an intelligent data display interface, including fault prediction, diagnosis and alarm mechanisms, thereby building a digital twin system; S4. Based on appropriate communication technologies, upload the information collected by sensors and cameras to the digital twin system; S5. Use inspection robots and cameras to collect power station data and complete the power station inspection.

[0006] Furthermore, the sensors in S1 include a temperature sensor, a pressure sensor, a current sensor, and a voltage sensor.

[0007] Furthermore, in S2, the areas and key equipment that need to be monitored are determined based on the power plant layout and equipment distribution. By simulating the monitoring range under different camera positions and angles, the coverage and resolution of each position are calculated. The camera positions and angles are adjusted to achieve the optimal balance between coverage and resolution. At the same time, the optimal camera deployment scheme is determined by considering camera cost and installation difficulty.

[0008] Furthermore, the digital twin power plant representation model in S3 utilizes laser scanning technology to perform a comprehensive scan of the power plant, combined with photogrammetry technology to take photos of the power plant from different angles using cameras deployed in S2, to construct a three-dimensional digital model of the power plant with layered display capabilities and support for multi-angle operation.

[0009] Furthermore, the digital twin system in S3 adopts a combination of rule-based reasoning and machine learning to establish a fault diagnosis rule base. It formulates fault judgment rules based on sensor data and image information features, uses machine learning algorithms to train historical inspection data to establish a fault prediction model, and compares and analyzes the real-time collected data with the fault diagnosis rules and prediction model during the inspection process to detect equipment faults and generate fault diagnosis reports. When the real-time collected monitoring data exceeds the set range, the system activates an abnormal alarm mechanism.

[0010] Furthermore, S4 selects LoRa wireless communication technology to transmit the data collected by the sensors and the image data captured by the camera to the central control system. The central control system performs real-time analysis and processing of the data, including using real-time data filtering algorithms to remove noise interference, using data compression technology to reduce the amount of data transmission and storage requirements, establishing an efficient data transmission channel, ensuring that real-time monitoring data can be transmitted to the digital twin model in real time and quickly, and accurately integrating the monitoring data into the three-dimensional model through data mapping algorithms and interpolation algorithms to achieve dynamic data updates.

[0011] Furthermore, the S5 inspection robot is equipped with multiple sensors and cameras, employing autonomous navigation technology. It achieves precise positioning and path planning of the power plant environment through a combination of LiDAR and inertial navigation systems. The path planning algorithm adopts an ant colony optimization strategy, establishing a map model of the inspection area based on the power plant layout and equipment distribution. The optimal inspection path is found by ants selecting paths and releasing pheromones. Different weights are assigned to different areas considering equipment importance and failure probability factors. Parallel computing technology is used to improve the real-time performance of path planning. During navigation, it perceives changes in the surrounding environment in real time and transmits the perceived information back to the digital twin system. The camera captures digital images within its range within a specified period and uploads them to the digital twin system in real time, completing timely system updates.

[0012] Beneficial Effects: Compared with existing technologies, the significant advantages of this invention are as follows: Through intelligent inspection robots and optimized path planning, power plant inspection tasks can be completed quickly and accurately, greatly improving inspection efficiency; advanced data analysis and fault diagnosis technologies can promptly detect equipment faults and provide detailed fault diagnosis reports, offering accurate guidance to maintenance personnel; real-time data monitoring and anomaly alarm functions can promptly detect abnormal situations in power plant operation, reminding staff to handle them in a timely manner, effectively improving power plant safety; digital twin modeling technology can visualize the power plant's operating status, facilitating management and decision-making by staff; the automated inspection system can reduce the workload of manual inspections and lower operating costs. Simultaneously, the optimized camera deployment scheme can also reduce equipment investment costs. Detailed Implementation

[0013] The technical solution of the present invention will be described in detail below with reference to specific examples.

[0014] A digital twin-driven method for inspecting new energy power plants includes the following steps: Step 1: Select different types of sensors that meet the requirements for different devices in the power station, and carry out comprehensive sensor installation throughout the power station to provide a simulated digital foundation for the digital twin system. Sensor selection refers to the precise placement of various high-performance sensors, including temperature sensors, pressure sensors, current sensors, and voltage sensors, at key equipment such as solar panels, inverters, and transformers, as well as key locations such as cable connections and switchgear, to achieve data acquisition within specific measurement accuracy ranges. During the deployment process, the sensor installation locations should accurately reflect the equipment operating status and achieve a measurement accuracy of ±0.5℃ within a temperature range of -40℃ to 125℃. Strain gauge pressure sensors are selected, which can achieve a measurement accuracy of ±0.1% FS under different pressure environments. Hall effect sensors are used for current sensors, which can accurately measure currents from milliamperes to kiloamperes with a measurement accuracy of ±0.5%. Voltage divider sensors are used for voltage sensors, which can achieve a measurement accuracy of ±0.2% at different voltage levels.

[0015] Step 2: For different locations within the power plant, use algorithms based on coverage and resolution optimization to plan and select suitable camera placement locations, providing a simulated image foundation for the digital twin system. Coverage refers to the ratio of the area covered by the cameras to the total monitored area, while resolution refers to the image clarity and detail rendering capability of the cameras at different locations for key equipment. In the specific implementation process, the areas to be monitored and key equipment need to be determined based on the power plant layout and equipment distribution. By simulating the monitoring range under different camera positions and angles, the coverage and resolution of each location are calculated. The camera positions and angles are adjusted to achieve an optimal balance between coverage and resolution. Simultaneously, factors such as camera cost and installation difficulty are considered to determine the optimal camera placement scheme.

[0016] Step 3: Externally, a digital twin power plant performance model is created by combining laser scanning and camera imaging; internally, an intelligent data display interface, fault prediction, diagnosis, and alarm mechanisms are formed using rule-based reasoning and machine learning, thereby building a digital twin system. In the specific implementation process, high-precision laser scanning technology is used to conduct a comprehensive scan of the power station. Combined with advanced photogrammetry technology, cameras deployed in step 2 are used to take photos of the power station from different angles, constructing a three-dimensional digital model of the power station with layered display capabilities and multi-angle operation support. Fault diagnosis adopts a method combining rule-based reasoning and machine learning. A fault diagnosis rule base is established, and fault judgment rules are formulated based on sensor data and image information features. Machine learning algorithms are used to train historical inspection data to build a fault prediction model. During the inspection process, the real-time collected data is compared and analyzed with the fault diagnosis rules and prediction model to promptly detect equipment faults and generate detailed fault diagnosis reports. When the monitored data exceeds the set range, the system immediately activates the abnormal alarm mechanism. Alarm methods include various forms such as audible and visual alarms, SMS notifications, and email notifications. The alarm information includes detailed abnormal data information, and alarms are graded according to the severity of the abnormality. Based on the above, a digital twin model is built, and multi-physics simulation functions are developed. Machine learning algorithms are used to analyze the historical operating data of the power station, establish equipment performance prediction models, and conduct fault simulation to provide decision support for fault handling.

[0017] Step 4: Select appropriate communication technology to complete the communication work for uploading information collected by sensors or cameras to the digital twin system, achieving seamless integration between the power plant and the digital twin system. In the specific implementation process, advanced LoRa wireless communication technology is selected to transmit data collected by sensors and images captured by cameras to the central control system. This system uses a high-performance data processing server equipped with professional data processing software to perform real-time analysis and processing of the data, including using real-time data filtering algorithms to remove noise interference and utilizing data compression technology to reduce data transmission volume and storage requirements. An efficient data transmission channel is established to ensure that real-time monitoring data can be transmitted to the digital twin model quickly and in real time. Through data mapping and interpolation algorithms, the monitoring data is accurately integrated into the 3D model, achieving dynamic data updates.

[0018] Step 5: Regularly use the intelligent inspection robot and the cameras deployed in Step 2 to collect internal and external data of the power station, completing routine power station inspections. The intelligent inspection robot is equipped with multiple high-precision sensors and high-resolution cameras, employing autonomous navigation technology. It combines LiDAR and inertial navigation systems to achieve precise positioning and path planning of the power station environment. The path planning algorithm uses an ant colony optimization strategy, establishing a map model of the inspection area based on the power station layout and equipment distribution. The optimal inspection path is found by using ants to select paths and release pheromones. Different weights are assigned to different areas considering factors such as equipment importance and failure probability. Parallel computing technology is used to improve the real-time performance of path planning. During navigation, it can perceive changes in the surrounding environment in real time and transmit the perceived information back to the digital twin system. The cameras capture digital images within their designated area within a specified period and upload them to the digital twin system in real time, ensuring timely system updates.

Claims

1. A digital twin driven new energy power plant inspection method, characterized in that, Includes the following steps: S1. Select and install various types of sensors at key locations of critical equipment in the power plant; S2. For different locations in the power station, use algorithms based on coverage and resolution optimization to plan and select suitable camera deployment locations. S3. By combining laser scanning and camera imaging, a digital twin power plant performance model is completed; rule-based reasoning and machine learning are used to form an intelligent data display interface, including fault prediction, diagnosis and alarm mechanisms, thereby building a digital twin system; S4. Based on appropriate communication technologies, upload the information collected by sensors and cameras to the digital twin system; S5. Use inspection robots and cameras to collect power station data and complete the power station inspection.

2. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, The sensors in S1 include a temperature sensor, a pressure sensor, a current sensor, and a voltage sensor.

3. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, In S2, the areas and key equipment that need to be monitored are determined based on the power plant layout and equipment distribution. The coverage and resolution of each location are calculated by simulating the monitoring range under different camera positions and angles. The camera positions and angles are adjusted to achieve the optimal balance between coverage and resolution. At the same time, the optimal camera deployment scheme is determined by considering camera cost and installation difficulty.

4. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, The digital twin power plant representation model in S3 utilizes laser scanning technology to perform a comprehensive scan of the power plant, combined with photogrammetry technology to take photos of the power plant from different angles using cameras deployed in S2, constructing a three-dimensional digital model of the power plant with layered display capabilities and supporting multi-angle operation.

5. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, The digital twin system in S3 adopts a combination of rule-based reasoning and machine learning to establish a fault diagnosis rule base. It formulates fault judgment rules based on sensor data and image information features, and uses machine learning algorithms to train historical inspection data to build a fault prediction model. During the inspection process, it compares and analyzes the real-time collected data with the fault diagnosis rules and prediction model to detect equipment faults and generate fault diagnosis reports. When the real-time collected monitoring data exceeds the set range, the system activates an abnormal alarm mechanism.

6. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, S4 selects LoRa wireless communication technology to transmit data collected by sensors and images captured by cameras to the central control system. The central control system performs real-time analysis and processing of the data, including using real-time data filtering algorithms to remove noise interference, using data compression technology to reduce data transmission volume and storage requirements, establishing an efficient data transmission channel, ensuring that real-time monitoring data can be transmitted to the digital twin model in real time and quickly, and accurately integrating the monitoring data into the three-dimensional model through data mapping and interpolation algorithms to achieve dynamic data updates.

7. The digital twin-driven inspection method for new energy power plants according to claim 1, characterized in that, The S5 inspection robot is equipped with multiple sensors and cameras, employing autonomous navigation technology. It achieves precise positioning and path planning of the power plant environment through a combination of LiDAR and inertial navigation systems. The path planning algorithm uses an ant colony optimization strategy, establishing a map model of the inspection area based on the power plant layout and equipment distribution. The optimal inspection path is found by ants selecting paths and releasing pheromones. Different weights are assigned to different areas considering equipment importance and failure probability factors. Parallel computing technology is used to improve the real-time performance of path planning. During navigation, it perceives changes in the surrounding environment in real time and transmits the perceived information back to the digital twin system. The camera captures digital images within its range within a specified period and uploads them to the digital twin system in real time, completing timely system updates.