Online intelligent monitoring system for thermal power plant

Abstract

The invention discloses an online intelligent monitoring system for a thermal power plant, and the system comprises a front-end sensing module which comprises a plurality of high-resolution thermal imaging cameras, a multi-view visible light camera, and a plurality of high-precision sensors; the data transmission module adopts a redundant communication architecture combining wired optical fibers and wireless 5G, and ensures that the data acquired by the front-end sensing module is stably transmitted to the background processing module at a high speed; the front-end sensing module comprises a high-resolution thermal imaging camera, a multi-view visible light camera and a plurality of high-precision sensors, the background processing module comprises an intelligent analysis server and a data storage server, and the system relates to the technical field of thermal power plant safety monitoring. The thermal power plant key equipment is monitored in real time in all directions, and in combination with an intelligent image recognition algorithm, subtle changes of equipment appearance, indicator lights and instrument states can be accurately recognized, so that comprehensive and accurate data support is provided for equipment operation and maintenance, and the risks of missing detection and misjudgment are effectively reduced.

Description

Technical Field

[0001] The invention relates to the technical field of thermal power plant safety monitoring, and in particular to an online intelligent monitoring system for a thermal power plant. Background Art

[0002] In the operation of thermal power plants, the stable operation of power equipment is crucial. Traditional equipment monitoring methods mainly rely on manual inspections. Operation and maintenance personnel need to regularly go to the site to inspect key equipment such as generators, transformers, switch cabinets, etc., observe the appearance of the equipment with the naked eye, use handheld thermometers to measure the temperature, and check the instrument readings. This method has many disadvantages: on the one hand, the efficiency of manual inspections is low. Faced with the huge and complex equipment group of thermal power plants, an inspection often consumes a lot of manpower and time, and it is difficult to achieve high-frequency inspections; on the other hand, manual inspections are greatly affected by subjective factors, and are prone to missed inspections and misjudgments. For example, some subtle deformations of the appearance of equipment and abnormal flashing of indicator lights may be difficult to detect. The measurement accuracy and selection of measurement positions of handheld thermometers will also vary from person to person, resulting in the discovery of equipment heating hazards that are not timely and accurate enough. In addition, manual inspections cannot grasp the operating status of the equipment in real time and in all aspects. During the interval between two inspections, if the equipment suddenly fails, it is difficult to respond in time, which greatly affects the safe production and power supply stability of the thermal power plant.

[0003] With the advancement of science and technology, some thermal power plants have introduced some automated monitoring methods, but there are still shortcomings. Some early monitoring systems can only achieve a single function, such as simple temperature monitoring or simple video monitoring, and lack the comprehensive intelligent monitoring capabilities of multiple technologies. In addition, in terms of data transmission, due to the limitations of communication technology, data delays and losses often occur, and real-time requirements cannot be met; in the background analysis and processing, the algorithm is simple, and it is difficult to conduct in-depth mining of massive equipment data, and it is impossible to accurately diagnose equipment defects and warn of failures in advance, which is difficult to meet the needs of efficient and safe operation of modern thermal power plants. Summary of the invention

[0004] The purpose of the present invention is to provide an online intelligent monitoring system for a thermal power plant, aiming to solve the above-mentioned problems existing in the prior art.

[0005] The technical solution of the present invention to achieve the above-mentioned purpose is an online intelligent monitoring system for a thermal power plant, comprising:

[0006] Front-end perception module, including multiple high-resolution thermal imaging cameras, multi-view visible light cameras, and a variety of high-precision sensors;

[0007] The thermal imaging cameras are distributed at key heating parts of thermal power plant equipment, capturing infrared radiation from the equipment surface in real time in a non-contact manner and converting it into temperature data. The visible light cameras fully cover the operating area of ​​the power equipment, with a resolution of no less than 1920×1080 pixels. The sensors include temperature sensors, pressure sensors, vibration sensors, and gas leakage sensors, all of which are deployed at various working points in the thermal power plant.

[0008] The data transmission module uses a redundant communication architecture that combines wired optical fiber and wireless 5G to ensure that the data collected by the front-end sensing module is stably and quickly transmitted to the back-end processing module;

[0009] The backend processing module includes an intelligent analysis server and a data storage server. The intelligent analysis server runs an AI algorithm based on deep learning, which is used to perform real-time analysis of front-end data to achieve equipment defect diagnosis and fault warning. The A1 algorithm includes target detection, image classification, and time series prediction. The data storage server is used to meet data storage needs and supports accurate retrieval of historical data, equipment ledgers, and inspection records.

[0010] Preferably, it also includes:

[0011] A temperature monitoring processing unit matched with the thermal imaging camera, the temperature monitoring processing unit is connected to the thermal imaging camera via Gigabit Ethernet, and has a built-in dynamically updateable normal operating temperature database of thermal power plant equipment, covering the temperature thresholds of at least 50 common equipment and 100 working modes. Once the received temperature data exceeds the threshold, the sound and light alarm is triggered within 1 second, and an alarm notification is pushed to the mobile phone text message, work email, and dedicated operation and maintenance APP of the operation and maintenance personnel, and the abnormal temperature point and the equipment where it is located are highlighted with a red flashing box on the monitoring software interface;

[0012] The temperature trend analysis module is connected to the temperature monitoring processing unit, uses a fitting algorithm to analyze the temperature trend, and automatically adjusts the analysis period according to the temperature fluctuation amplitude of the equipment.

[0013] Preferably, it also includes an intelligent analysis server connected to the multi-view visible light camera, which has a built-in deep learning image recognition algorithm that has been trained and optimized with no less than 100,000 sample images. For different power equipment, it can accurately identify subtle deformations in the appearance, 8 color states and 5 flashing frequencies of indicator lights. When it is determined that an error or abnormality has occurred in the equipment, an alarm is triggered within 3 seconds, the discovery time and equipment location are recorded, and an alarm log is generated.

[0014] Preferably, it also includes an image preliminary screening module, which captures images regularly through multi-angle visible light cameras and high-resolution thermal imaging cameras distributed around the equipment. The image preliminary screening module automatically marks suspected abnormal images and pushes them to the manual review end through convolutional neural network image recognition technology;

[0015] The manual review terminal is equipped with a 4K high-definition display screen and is placed in the monitoring center. The operation and inspection personnel review the monthly updated equipment appearance standard library and temperature threshold range criteria to generate inspection records and inspection reports.

[0016] Preferably, the inspection record includes the electronic signature of the inspector, the precise time of review, a graphic description of the abnormal details, and handling opinions.

[0017] Preferably, the alarm log includes abnormal image snapshots and abnormal detailed descriptions.

[0018] Preferably, the suspected abnormal images include three types: appearance abnormality, temperature abnormality and component displacement.

[0019] Preferably, the temperature trend analysis module can export the temperature trend data in .csv format, which is convenient for docking with third-party data analysis software.

[0020] The beneficial effects of the present invention are:

[0021] 1. This system realizes all-round real-time monitoring of key equipment in thermal power plants through the coordinated work of high-resolution thermal imaging cameras, multi-view visible light cameras and a variety of high-precision sensors in the front-end perception module. The thermal imaging camera uses a highly sensitive infrared detector to accurately capture the temperature of key heating parts of the equipment with a measurement accuracy of ±2°C. With the adaptively adjusted measurement frequency, it can detect temperature anomalies in a timely manner. The visible light camera covers the equipment operating area with high pixels, and combined with the intelligent image recognition algorithm, it can accurately identify subtle changes in the appearance, indicator lights and instrument status of the equipment, providing comprehensive and accurate data support for equipment operation and maintenance, effectively reducing the risk of missed detection and misjudgment.

[0022] 2. A redundant communication architecture combining wired optical fiber and wireless 5G is adopted to ensure high-speed and stable data transmission. In particular, thermal imaging data and alarm information are transmitted through the 5G network with a delay of no more than 10 milliseconds to ensure real-time performance. The intelligent analysis server of the background processing module runs advanced AI algorithms and can quickly process large amounts of image and sensor data to achieve equipment defect diagnosis and fault warning, greatly improving operation and maintenance efficiency, reducing downtime caused by equipment failures, and ensuring a continuous and stable power supply.

[0023] 3. The integration of AI intelligent visible light inspection and thermal imaging inspection and analysis units can not only automatically identify abnormalities of various equipment, but also compare and analyze thermal images in real time during the inspection process, generate detailed thermal fault hazard reports, and provide targeted processing suggestions for operation and maintenance personnel. The online manual image inspection function is used as a supplement to adapt to special working conditions, ensure uninterrupted inspections, effectively improve emergency response speed, and reduce accident losses. DETAILED DESCRIPTION

[0024] 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.

[0025] The present invention provides a technical solution: an online intelligent monitoring system for a thermal power plant, comprising:

[0026] Front-end perception module, including multiple high-resolution thermal imaging cameras, multi-view visible light cameras, and a variety of high-precision sensors;

[0027] The thermal imaging cameras are distributed at key heating parts of thermal power plant equipment, capturing infrared radiation from the equipment surface in real time in a non-contact manner and converting it into temperature data. The visible light cameras fully cover the operating area of ​​the power equipment, with a resolution of no less than 1920×1080 pixels. The sensors include temperature sensors, pressure sensors, vibration sensors, and gas leakage sensors, all of which are deployed at various working points in the thermal power plant.

[0028] The data transmission module uses a redundant communication architecture that combines wired optical fiber and wireless 5G to ensure that the data collected by the front-end sensing module is stably and quickly transmitted to the back-end processing module;

[0029] The backend processing module includes an intelligent analysis server and a data storage server. The intelligent analysis server runs an AI algorithm based on deep learning, which is used to perform real-time analysis of front-end data to achieve equipment defect diagnosis and fault warning. The A1 algorithm includes target detection, image classification, and time series prediction. The data storage server is used to meet data storage needs and supports accurate retrieval of historical data, equipment ledgers, and inspection records.

[0030] This embodiment is further configured to include:

[0031] A temperature monitoring processing unit matched with the thermal imaging camera, the temperature monitoring processing unit is connected to the thermal imaging camera via Gigabit Ethernet, and has a built-in dynamically updateable normal operating temperature database of thermal power plant equipment, covering the temperature thresholds of at least 50 common equipment and 100 working modes. Once the received temperature data exceeds the threshold, the sound and light alarm is triggered within 1 second, and an alarm notification is pushed to the mobile phone text message, work email, and dedicated operation and maintenance APP of the operation and maintenance personnel, and the abnormal temperature point and the equipment where it is located are highlighted with a red flashing box on the monitoring software interface;

[0032] The temperature trend analysis module is connected to the temperature monitoring processing unit, uses a fitting algorithm to analyze the temperature trend, and automatically adjusts the analysis period according to the temperature fluctuation amplitude of the equipment.

[0033] This embodiment is further configured to also include an intelligent analysis server connected to the multi-view visible light camera, which has a built-in deep learning image recognition algorithm that has been trained and optimized with no less than 100,000 sample images. For different power equipment, the server can accurately identify subtle deformations in the appearance, 8 color states and 5 flashing frequencies of indicator lights. When it is determined that an error or abnormality has occurred in the equipment, an alarm is triggered within 3 seconds, the discovery time and equipment location are recorded, and an alarm log is generated.

[0034] This embodiment is further configured to include an image preliminary screening module, which captures images regularly through multi-angle visible light cameras and high-resolution thermal imaging cameras distributed around the device. The image preliminary screening module automatically marks suspected abnormal images and pushes them to the manual review end through convolutional neural network image recognition technology;

[0035] The manual review terminal is equipped with a 4K high-definition display screen and is placed in the monitoring center. The operation and inspection personnel review the monthly updated equipment appearance standard library and temperature threshold range criteria to generate inspection records and inspection reports.

[0036] This embodiment is further configured such that the inspection record includes the electronic signature of the inspection personnel, the precise time of review, a graphic description of the abnormal details, and handling opinions.

[0037] This embodiment is further configured such that the alarm log includes an abnormal image snapshot and a detailed description of the abnormality.

[0038] This embodiment is further configured such that the suspected abnormal images include three types: appearance abnormality, temperature abnormality and component displacement.

[0039] The present embodiment is further configured such that the temperature trend analysis module can export the temperature trend data in .csv format, so as to facilitate connection with third-party data analysis software.

[0040] The present invention will be further described below in conjunction with the embodiments.

[0041] Example:

[0042] This system is based on the principle of infrared thermal imaging. The infrared detector in the thermal imaging camera receives infrared radiation emitted by the surface of the device. According to the blackbody radiation law, the higher the temperature of the object, the stronger the infrared energy it radiates. The detector converts the infrared energy into an electrical signal, and then obtains the temperature data of the device surface through signal processing and conversion. By comparing the preset temperature threshold in real time, it can be judged whether the device is abnormally hot.

[0043] In terms of image recognition, a deep learning algorithm is used to build a model, and a large number of image samples containing normal and abnormal equipment status are trained to enable the model to grasp the characteristic patterns of different equipment in terms of appearance, indicator lights, instrument readings, etc. When real-time visible light images are input, the model can quickly identify the differences from the normal pattern, thereby determining whether there is any abnormality in the equipment.

[0044] During the data transmission and processing process, wired optical fiber provides a stable and reliable basic transmission link, and the wireless 5G network uses its high bandwidth and low latency characteristics to ensure instant data transmission for data with extremely high real-time requirements such as thermal imaging. The background processing module integrates and processes the multi-source heterogeneous data collected by the front-end perception module based on big data analysis and AI algorithms, and mines the equipment operation status information hidden behind the data to achieve early warning and accurate diagnosis of equipment failures, thereby ensuring the safe and efficient operation of the power equipment in the thermal power plant.

[0045] The specific workflow is:

[0046] 1. Data collection: The thermal imaging camera of the front-end perception module captures the infrared radiation of the key heating parts of the key equipment of the thermal power plant in real time according to the preset distribution position, and converts it into temperature data. At the same time, the visible light camera collects image information of the equipment operation area in all directions. Various high-precision sensors synchronously measure temperature, pressure, vibration, gas leakage and other parameters. These data are stably transmitted to the background processing module through wired optical fiber or wireless 5G network according to the set transmission protocol and encryption method.

[0047] 2. Real-time monitoring and alarm: The temperature monitoring processing unit in the background receives the temperature data from the thermal imaging camera and compares it with the built-in dynamically updated normal operating temperature database of thermal power plant equipment. Once the temperature exceeds the threshold, the sound and light alarm is triggered immediately within 1 second, and the alarm notification is pushed to the operation and maintenance personnel through mobile phone text messages, work emails, and dedicated operation and maintenance APP. At the same time, the abnormal temperature point and the equipment where it is located are highlighted in a red flashing box on the monitoring software interface with a flashing frequency of 2Hz. The intelligent analysis server analyzes the visible light camera image and sensor data in real time. When abnormalities such as equipment appearance, indicator lights, and instrument readings are identified, an alarm is also quickly triggered and detailed abnormal information is recorded.

[0048] 3. Inspection process:

[0049] AI intelligent visible light inspection: The front-end multi-view visible light camera continuously collects images and transmits them to the intelligent analysis server. The server has a built-in deep learning image recognition algorithm that has been optimized through training with a large number of sample images. It analyzes images in real time. For different power equipment, it can accurately identify subtle deformations of the appearance (with an accuracy of 1 mm deformation), 8 color states and 5 flashing frequencies of indicator lights, and instrument readings accurate to two decimal places. When it is judged that the transformer oil level deviation exceeds the normal range of ±5%, the circuit breaker opening and closing indication is wrong, and other abnormalities, an alarm is triggered within 3 seconds, and the discovery time is recorded accurately to milliseconds and the equipment location is accurate to centimeters. An alarm log containing abnormal image snapshots and abnormal detailed descriptions is generated.

[0050] Thermal imaging inspection and analysis: The thermal imaging inspection and analysis unit combines the thermal images collected by the thermal imaging camera with AI technology. During the inspection process, it compares the real-time thermal images of the equipment with the pre-stored high-resolution thermal image templates of the same working conditions. It uses the image segmentation algorithm to automatically select the abnormal heating area and zoom in on the thermal image details with a magnification of at least 4 times. It automatically generates a thermal fault hazard report, which covers the description of the problem location accurate to 1 square centimeter and the estimated equipment list within a 5-meter radius of the fault impact.

[0051] Online manual image inspection: The system takes pictures regularly according to the preset inspection plan through visible light cameras and infrared cameras distributed around the equipment. The picture resolution is 1080P and the frame rate is 25 frames per second. The plan can be flexibly adjusted according to the equipment maintenance cycle, seasonal temperature and humidity, and real-time load rate. The image preliminary screening module uses convolutional neural network image recognition technology with a screening accuracy of 90%. Suspected abnormal pictures are automatically labeled according to three categories: abnormal appearance, abnormal temperature, and component displacement and pushed to the manual review link. The manual review terminal is equipped with a 27-inch 4K high-definition display and placed in the monitoring center. The operation and inspection personnel review the monthly updated equipment appearance standard library, temperature threshold range and other criteria, and generate inspection records containing the inspector's electronic signature, precise review time, graphic description of abnormal details, and handling opinions. Inspection reports compatible with Microsoft Word and Excel can be exported.

[0052] 4. Data storage and analysis: The data storage server uses RAID 5 redundant array technology to centrally store historical data, equipment ledgers, inspection records, etc. The storage capacity meets the 10-year operating data storage needs of the thermal power plant, and supports accurate retrieval of historical data, equipment ledgers, and inspection records within 5 seconds, with a retrieval accuracy of 99%. The intelligent analysis server uses deep learning-based target detection, image classification, and time series prediction AI algorithms to deeply mine the massive data collected by the front end, continuously optimize the equipment defect diagnosis model and fault warning mechanism, and provide long-term data support and decision-making basis for equipment operation and maintenance.

[0053] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. An online intelligent monitoring system for a thermal power plant, characterized in that: include: Front-end perception module, including multiple high-resolution thermal imaging cameras, multi-view visible light cameras, and a variety of high-precision sensors; The thermal imaging cameras are distributed at key heating parts of thermal power plant equipment, capturing infrared radiation from the equipment surface in real time in a non-contact manner and converting it into temperature data. The visible light cameras fully cover the operating area of ​​the power equipment, with a resolution of no less than 1920×1080 pixels. The sensors include temperature sensors, pressure sensors, vibration sensors, and gas leakage sensors, all of which are deployed at various working points in the thermal power plant. The data transmission module uses a redundant communication architecture that combines wired optical fiber and wireless 5G to ensure that the data collected by the front-end sensing module is stably and quickly transmitted to the back-end processing module; The backend processing module includes an intelligent analysis server and a data storage server. The intelligent analysis server runs an AI algorithm based on deep learning, which is used to perform real-time analysis of front-end data to achieve equipment defect diagnosis and fault warning. The A1 algorithm includes target detection, image classification, and time series prediction. The data storage server is used to meet data storage needs and supports accurate retrieval of historical data, equipment ledgers, and inspection records.

2. The online intelligent monitoring system for a thermal power plant according to claim 1 is characterized in that: Also includes: A temperature monitoring processing unit matched with the thermal imaging camera, the temperature monitoring processing unit is connected to the thermal imaging camera via Gigabit Ethernet, and has a built-in dynamically updateable normal operating temperature database of thermal power plant equipment, covering the temperature thresholds of at least 50 common equipment and 100 working modes. Once the received temperature data exceeds the threshold, the sound and light alarm is triggered within 1 second, and an alarm notification is pushed to the mobile phone text message, work email, and dedicated operation and maintenance APP of the operation and maintenance personnel, and the abnormal temperature point and the equipment where it is located are highlighted with a red flashing box on the monitoring software interface; The temperature trend analysis module is connected to the temperature monitoring processing unit, uses a fitting algorithm to analyze the temperature trend, and automatically adjusts the analysis period according to the temperature fluctuation amplitude of the equipment.

3. The online intelligent monitoring system for a thermal power plant according to claim 2 is characterized in that: It also includes an intelligent analysis server connected to the multi-view visible light camera, which has a built-in deep learning image recognition algorithm that has been trained and optimized with no less than 100,000 sample images. For different power equipment, it can accurately identify subtle deformations in the appearance, 8 color states and 5 flashing frequencies of indicator lights. When it is determined that an error or abnormality has occurred in the equipment, an alarm is triggered within 3 seconds, recording the discovery time and equipment location, and generating an alarm log.

4. The online intelligent monitoring system for a thermal power plant according to claim 3 is characterized in that: It also includes an image preliminary screening module, which uses multi-angle visible light cameras and high-resolution thermal imaging cameras distributed around the equipment to capture images at regular intervals. The image preliminary screening module uses convolutional neural network image recognition technology to automatically mark suspected abnormal images and push them to the manual review end; The manual review terminal is equipped with a 4K high-definition display screen and is placed in the monitoring center. The operation and inspection personnel review the monthly updated equipment appearance standard library and temperature threshold range criteria to generate inspection records and inspection reports.

5. The online intelligent monitoring system for a thermal power plant according to claim 4 is characterized in that: The inspection record includes the electronic signature of the inspector, the precise time of review, a graphic description of the abnormal details, and handling opinions.

6. The online intelligent monitoring system for a thermal power plant according to claim 3 is characterized in that: The alarm log includes an abnormal image snapshot and a detailed description of the abnormality.

7. The online intelligent monitoring system for a thermal power plant according to claim 4 is characterized in that: The suspected abnormal images include three types: appearance abnormality, temperature abnormality and component displacement.

8. The online intelligent monitoring system for a thermal power plant according to claim 2 is characterized in that: The temperature trend analysis module can export temperature trend data in .csv format, which is convenient for docking with third-party data analysis software.