An intelligent monitoring system and method for a thermal power boiler

Abstract

The present application relates to the technical field of automatic monitoring of thermal power boilers, and particularly relates to a thermal power boiler intelligent monitoring system and method, comprising a boiler intelligent monitoring terminal, a multi-dimensional correlation control device, a hierarchical response terminal and a cloud collaborative analysis platform; the boiler intelligent monitoring terminal is used to realize local collaborative analysis and double-condition triggered early warning; the multi-dimensional correlation control device is linked with the boiler intelligent monitoring terminal to collect data and execute control instructions, and simultaneously generates hierarchical response instructions; the hierarchical response terminal adopts a four-level response plus automatic upgrade mechanism to ensure accurate and rapid disposal; the cloud platform is used to generate a four-in-one collaborative optimization strategy of safety, energy efficiency, environmental protection and maintenance and to issue iterations, realizing an intelligent operation and maintenance mode change from passive response to active prediction, and solving the problems of single monitoring dimension, independent control subsystems, delayed early warning response and lack of multi-target collaborative optimization in the prior art.

Description

Technical Field

[0001] This invention relates to the field of automated monitoring technology for thermal power boilers, specifically to an intelligent monitoring system and method for thermal power boilers. Background Technology

[0002] The thermal power generation industry is facing unprecedented pressure and challenges in its transformation. On the one hand, as a stable base load of the power system, thermal power generation will continue to play a core role in ensuring energy security for a considerable period of time. On the other hand, increasingly stringent requirements from the whole society for environmental protection and energy conservation and emission reduction are driving thermal power companies to develop in a cleaner, lower-carbon, and more efficient direction. Against this macro-background, thermal power companies must not only ensure the absolute safety and stable operation of boilers—the core equipment—avoiding any form of unplanned shutdowns and major safety accidents, but also continuously improve combustion efficiency to reduce coal consumption, strictly control the emission intensity of pollutants such as carbon dioxide, sulfur oxides, and nitrogen oxides, while extending equipment lifespan to reduce total life-cycle operating costs. However, these four objectives—safety, energy efficiency, environmental protection, and maintenance—often present contradictions and trade-offs in actual operation. Therefore, how to construct an intelligent monitoring and control system that can comprehensively consider these four core objectives and achieve overall optimization has become a core challenge that urgently needs to be overcome in the field of thermal power technology.

[0003] Currently, the conventional monitoring and control technology system for thermal power boilers exhibits significant functional fragmentation and performance limitations when addressing the aforementioned complex and multi-objective operational requirements. At the monitoring level, traditional systems heavily rely on distributed conventional thermal instruments such as thermocouples, pressure transmitters, and flow meters. These instruments primarily collect limited physical parameters such as temperature, pressure, flow rate, and water level. This point-based measurement approach fails to detect crucial status information such as the three-dimensional distribution of the combustion field within the furnace, the shift of the flame center, the early evolution of coking on the water-cooled wall surface, and microscopic defects in the superheater / reheater tube panels caused by high-temperature creep or corrosion. For these latent anomalies, power plants typically rely on periodic manual inspections and experience-based judgment by operators. This is not only inefficient but also hinders comprehensive, real-time, and accurate perception due to the high temperature, high dust, and enclosed environment inside the furnace, preventing many potential faults from being detected in their early stages. At the control level, various subsystems, such as combustion optimization, flue gas system coordination, automatic water supply regulation, wastewater control, waste heat recovery, and environmental protection facilities like desulfurization, denitrification, and dust removal, are often provided by different suppliers, each following its own control logic, forming independently operating "information silos." The lack of effective data sharing and linkage mechanisms between these subsystems leads to a disconnect between combustion adjustments and pollutant generation control, a mismatch between load changes and water supply and air volume responses, and a lack of effective correlation between equipment maintenance decisions and real-time operational status. This fragmented control model prevents the system from taking a holistic approach and forming a coordinated control strategy targeting multiple objectives: safety, energy efficiency, environmental protection, and maintenance.

[0004] Due to the limitations of the aforementioned monitoring and control levels, existing early warning and emergency response mechanisms appear simplistic, rigid, and inefficient. Meanwhile, massive amounts of operational data remain locally, lacking effective mechanisms for organization, cleaning, and utilization. This prevents the use of the powerful computing and storage capabilities of the cloud for deep learning and model iteration, resulting in the waste of potential patterns and optimization knowledge contained within the data.

[0005] In summary, the inherent deficiencies in the existing technology system regarding comprehensive perception, collaborative control, intelligent early warning, and proactive maintenance have severely hampered further improvements in the operational level of thermal power boilers. Summary of the Invention

[0006] To address the problems of poor comprehensive perception and collaborative control capabilities in existing technologies for monitoring thermal power boilers, resulting in insufficient monitoring response and data analysis capabilities, this invention provides an intelligent monitoring system for thermal power boilers.

[0007] To achieve the above objectives, the present invention employs the following technical solution: This invention provides an intelligent monitoring system for thermal power boilers, including a boiler intelligent monitoring terminal, multi-dimensional correlation control equipment, hierarchical response terminal, and cloud-based collaborative analysis platform; The intelligent boiler monitoring terminal is communicatively connected to the multi-dimensional correlation control device and the cloud-based collaborative analysis platform, respectively. It is used to collect boiler operating status monitoring data and perform local collaborative analysis. Based on the local collaborative analysis results, it sends execution control commands to the multi-dimensional correlation control device and uploads the monitoring data to the cloud-based collaborative analysis platform. The multi-dimensional correlation control device is communicatively connected to the hierarchical response terminal and simultaneously to the boiler intelligent monitoring terminal. It is used to collect data in conjunction with the boiler intelligent monitoring terminal, execute the execution control command, generate early warning information and hierarchical response command based on the analysis results of the boiler intelligent monitoring terminal or the cloud collaborative analysis platform, and send the early warning information and hierarchical response command to the hierarchical response terminal. The tiered response terminal is used to receive the warning information and tiered response instructions, and to execute the corresponding tiered response operations; The cloud-based collaborative analysis platform is used to receive the monitoring data uploaded by the boiler intelligent monitoring terminal, analyze and process the monitoring data to generate optimization strategies, and then distribute the optimization strategies to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control device.

[0008] Optionally, the boiler intelligent monitoring terminal integrates a multi-dimensional parameter acquisition module and an AI visual monitoring module; the multi-dimensional parameter acquisition module is equipped with a high-precision thermocouple sensor, a pressure transmitter, a gas sensor, a liquid level sensor, an ultrasonic vibration sensor, and a vibration sensor; the AI ​​visual monitoring module uses a high-temperature resistant infrared high-definition camera and a lidar with a temperature resistance of not less than 1200℃ to collect image data of furnace flame morphology, furnace wall coking status, and tube screen surface defects.

[0009] Optionally, the boiler intelligent monitoring terminal has a built-in local collaborative analysis unit, which has a built-in lightweight fusion model. The lightweight fusion model has three levels of early warning thresholds: regular deviation, suspicious anomaly, and emergency fault. It adopts a dual-condition triggering mechanism that requires both parameter anomaly and visual feature anomaly to be met simultaneously. The local collaborative analysis unit has a full data storage capacity of at least 168 hours.

[0010] Optionally, the boiler intelligent monitoring terminal also includes a real-time early warning triggering module. The real-time early warning triggering module adopts a hierarchical early warning mechanism, wherein regular deviations trigger platform text prompts and parameter adjustment suggestions, suspicious abnormalities trigger audible and visual prompts and pre-push combustion optimization instructions, and emergency faults trigger strong light, voice alarms, cloud emergency early warnings, and activation of associated equipment protection modes.

[0011] Optionally, the multi-dimensional correlation control equipment includes a combustion optimization controller, an online carbon emission monitor, a waste heat recovery intelligent regulator, a furnace tube corrosion monitoring module, a feedwater system controller, an induced draft / forced draft regulator, a blowdown valve control module, and a boiler life cycle archive. The combustion optimization controller, carbon emission online monitor, waste heat recovery intelligent regulator, furnace tube corrosion monitoring module, water supply system controller, induced draft / forced draft regulator, and blowdown valve control module are all communicatively connected to the boiler intelligent monitoring terminal, and are linked with the boiler intelligent monitoring terminal to collect data and execute the execution control commands. The boiler lifecycle archive is used to store equipment lifecycle data for multi-dimensional comparison.

[0012] Optionally, the multi-dimensional correlation control device and the boiler intelligent monitoring terminal achieve data interoperability through dual-mode communication of 5G, industrial Ethernet and wireless communication.

[0013] Optionally, the tiered response terminal includes an operation and maintenance management platform, mobile phones of operation and maintenance personnel, a power plant DCS system, an environmental protection supervision platform, and an emergency response center.

[0014] Optionally, the tiered response terminal adopts a four-level response plus automatic upgrade mechanism, wherein the operation and maintenance management platform and the mobile phones of operation and maintenance personnel are configured as the first priority, with a response time of 0 to 45 seconds; the power plant DCS system is configured as the second priority, with a response time of 45 to 90 seconds; the environmental protection supervision platform is configured as the third priority, with a response time of 90 to 120 seconds; and the emergency response center is configured as the fourth priority, with a response time after 120 seconds.

[0015] Optionally, the cloud-based collaborative analysis platform includes a multi-source data fusion module, an AI visual recognition algorithm module, a carbon emission-combustion linkage algorithm module, a predictive maintenance model, a multi-dimensional linkage algorithm module, and a collaborative optimization engine; The multi-source data fusion module is used to align, clean, and distribute the acquired boiler operation status monitoring data to downstream modules; The AI ​​visual recognition algorithm module is used to receive aligned boiler operation status monitoring data from the multi-source data fusion module for identification, and generate a structured report containing fault type, location, confidence level, and timestamp for collaborative optimization engine decision-making; The carbon emission-combustion linkage algorithm module is used to construct a mapping relationship between fuel quantity, air volume, load, and emission concentration based on the aligned boiler operation status monitoring data received from the multi-source data fusion module. It then calculates in real time how to minimize CO2 emissions or reduce NO emissions under the current load. x The optimal air-coal ratio with the best overall thermal efficiency is determined, and early warnings are generated and incorporated into the collaborative optimization engine's optimization strategy. The predictive maintenance model is used to generate specific maintenance suggestions based on the aligned boiler operating status monitoring data received by the multi-source data fusion module and output them to the collaborative optimization engine. The collaborative optimization engine integrates information generated by the AI ​​visual recognition algorithm module, the carbon emission-combustion linkage algorithm module, the predictive maintenance model, and the multi-dimensional linkage algorithm module to generate optimization strategies.

[0016] The present invention also provides a monitoring and control method based on the above-mentioned intelligent monitoring system for thermal power boilers, comprising the following steps: Boiler intelligent monitoring terminal collects boiler operation status monitoring data, including multi-dimensional physical parameters and AI visual image data inside the furnace. The monitoring data is then analyzed locally and executed control commands are sent to the multi-dimensional associated control device. At the same time, the monitoring data is uploaded to the cloud collaborative analysis platform. The system utilizes multi-dimensional correlation control devices to collect data in conjunction with the intelligent boiler monitoring terminal, executes the execution control commands, and generates early warning information and graded response commands based on the analysis results of the intelligent boiler monitoring terminal or the cloud-based collaborative analysis platform, which are then sent to the graded response terminal. The system utilizes a tiered response terminal to receive the warning information and tiered response instructions, and executes the corresponding tiered response operations. The received monitoring data is analyzed and processed using a cloud-based collaborative analysis platform to generate a collaborative optimization strategy encompassing four dimensions: safety, energy efficiency, environmental protection, and maintenance. This optimization strategy is then distributed to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control equipment to iteratively update the control parameters.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides an intelligent monitoring system for thermal power generation boilers. By arranging an intelligent boiler monitoring terminal, the system can perform preliminary processing and judgment at the data source. Even when communication with the cloud is interrupted, it can still maintain basic monitoring and local control functions, greatly improving the robustness and real-time response capability of the system, and avoiding monitoring blind areas caused by network delay or interruption. Meanwhile, the terminal controls associated devices downward and transmits data upward, forming a local closed loop of data collection, analysis, execution and cloud uploading, which provides a high-quality data foundation for subsequent cloud optimization. By arranging multi-dimensional associated control devices in communication connection with the intelligent boiler monitoring terminal, scattered control devices are connected with the monitoring terminal and response terminals through a unified communication interface, enabling them to receive collaborative instructions from the terminal or the cloud, realize linkage execution and actively generate graded response instructions based on analysis results. This shares part of the computing pressure of the terminal and the cloud platform, and improves the overall response efficiency of the system. The existence of graded response terminals enables early warning information to be delivered accurately according to preset priorities and scope of responsibilities, thereby effectively avoiding information overload, ensuring that high-risk emergency events receive the highest attention, while low-risk events are not over-responded, thus optimizing the allocation of operation and maintenance resources. The arrangement of the cloud collaborative analysis platform breaks through the limitation of local computing capacity, can process massive multi-source heterogeneous data, run complex AI models, thereby generate non-linear and forward-looking optimization strategies, and realize dynamic iteration of strategies, endowing the system with self-learning and self-optimization capabilities. The system has a simple structure. Through the combination and collaboration of the intelligent boiler monitoring terminal, multi-dimensional associated control devices, graded response terminals and cloud collaborative analysis platform, it realizes a full-chain closed loop of data collection, local analysis, graded execution and early warning response, in-depth cloud optimization and strategy feedback, and fundamentally solves the systematic problems in the prior art such as single monitoring dimension, control islands, chaotic early warning and lack of self-optimization capability.

[0018] The present invention also provides a monitoring control method based on the above-mentioned intelligent monitoring system for thermal power generation boilers. In this method, the terminal executes local control immediately after collection and analysis, and meanwhile associated devices generate early warnings based on more comprehensive analysis results, which realizes the synchronization of rapid control and accurate early warning. The whole process forms a complete intelligent closed loop, and transforms isolated, passive, single-objective operations into global, active, multi-objective collaborative intelligent management. It fundamentally solves the problems of incomplete perception, uncoordinated control, inaccurate early warning and unsustainable optimization in the prior art, and realizes a global optimal monitoring strategy for the operation of thermal power boilers. Description of Drawings

[0019] Figure 1 is a schematic structural diagram of an intelligent monitoring system for a thermal power generation boiler according to the present invention. Figure 2 is a structural framework diagram of an internal module of the intelligent boiler monitoring terminal of the present invention; Figure 3 This is a structural framework diagram of the cloud-based collaborative analysis platform module of the present invention; Figure 4 This is a schematic diagram of the intelligent monitoring and control method for thermal power boilers according to the present invention. Figure 5 This is a flowchart of system data linkage and early warning response in a specific embodiment of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0023] See Figure 1 This invention provides an intelligent monitoring system for thermal power boilers, including a boiler intelligent monitoring terminal, multi-dimensional correlation control equipment, hierarchical response terminal, and cloud-based collaborative analysis platform; See Figure 2The boiler intelligent monitoring terminal is communicatively connected to a multi-dimensional associated control device and a cloud-based collaborative analysis platform. It features multi-dimensional parameter acquisition, AI visual furnace monitoring, local collaborative analysis, real-time early warning triggering, and device collaborative communication functions. It collects boiler operating status monitoring data and performs local collaborative analysis. Based on the local collaborative analysis results, it issues execution control commands to the multi-dimensional associated control device and uploads the monitoring data to the cloud-based collaborative analysis platform. Preferably, the AI ​​visual monitoring module uses a high-temperature resistant infrared high-definition camera and lidar with a temperature resistance of not less than 1200℃ to collect furnace flame morphology, furnace wall coking status, and tube screen surface defect image data. The boiler intelligent monitoring terminal has a built-in local collaborative analysis unit with a lightweight fusion model. This lightweight fusion model has three levels of early warning thresholds: normal deviation, suspicious anomaly, and emergency fault. It employs a dual-condition triggering mechanism where both parameter anomalies and visual feature anomalies must be simultaneously satisfied. Furthermore, the local collaborative analysis unit has at least 168 hours of full data storage capacity. Preferably, the real-time early warning triggering module adopts a hierarchical early warning mechanism, wherein regular deviations trigger platform text prompts and parameter adjustment suggestions, suspicious anomalies trigger audio-visual prompts and pre-push combustion optimization instructions, and emergency faults trigger strong light, voice alarms, cloud emergency early warnings, and activation of associated equipment protection modes.

[0024] The multi-dimensional correlation control device is communicatively connected to the hierarchical response terminal and simultaneously to the boiler intelligent monitoring terminal. It includes a combustion optimization controller, an online carbon emission monitor, a waste heat recovery intelligent regulator, a furnace tube corrosion monitoring module, a feedwater system controller, an induced draft / forced draft regulator, a blowdown valve control module, and a boiler lifecycle archive. This device is used to collect data in conjunction with the boiler intelligent monitoring terminal, execute the execution control commands, and generate early warning information and hierarchical response commands based on the analysis results from the boiler intelligent monitoring terminal or the cloud-based collaborative analysis platform. The early warning information and hierarchical response commands are then sent to the hierarchical response terminal. The combustion optimization controller, online carbon emission monitor, waste heat recovery intelligent regulator, furnace tube corrosion monitoring module, feedwater system controller, induced draft / forced draft regulator, and blowdown valve control module are all communicatively connected to the boiler intelligent monitoring terminal, collecting data and executing the execution control commands in conjunction with the boiler intelligent monitoring terminal. The boiler lifecycle archive stores equipment lifecycle data for multi-dimensional comparison.

[0025] The tiered response terminal includes an operation and maintenance management platform, mobile phones of operation and maintenance personnel, a power plant DCS system, an environmental protection supervision platform, and an emergency response center. It is used to receive the early warning information and tiered response instructions, and execute corresponding tiered response operations. The multi-dimensional associated control equipment and the boiler intelligent monitoring terminal achieve data interoperability through dual-mode communication of 5G, industrial Ethernet, and wireless communication. Preferably, the tiered response terminal adopts a four-level response plus automatic upgrade mechanism. The operation and maintenance management platform and mobile phones of operation and maintenance personnel are configured as the first priority, with a response time of 0 to 45 seconds; the power plant DCS system is configured as the second priority, with a response time of 45 to 90 seconds; the environmental protection supervision platform is configured as the third priority, with a response time of 90 to 120 seconds; and the emergency response center is configured as the fourth priority, with a response time after 120 seconds.

[0026] See Figure 3 The cloud-based collaborative analysis platform includes a multi-source data fusion module, an AI visual recognition algorithm module, a carbon emission-combustion linkage algorithm module, a predictive maintenance model, a multi-dimensional linkage algorithm module, and a collaborative optimization engine. It receives monitoring data uploaded by the boiler intelligent monitoring terminal, analyzes and processes the monitoring data to generate optimization strategies, and distributes the optimization strategies to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control devices. Specifically, the multi-source data fusion module aligns and cleans the acquired boiler operating status monitoring data and distributes it to downstream modules; the AI ​​visual recognition algorithm module receives aligned boiler operating status monitoring data from the multi-source data fusion module, identifies the data, and generates a structured report containing fault type, location, confidence level, and timestamp for the collaborative optimization engine to make decisions; the carbon emission-combustion linkage algorithm module constructs a mapping relationship between fuel quantity, air volume, load, and emission concentration based on the aligned boiler operating status monitoring data received from the multi-source data fusion module, and solves in real time how to minimize CO2 emissions or minimize NO emissions under the current load. x The optimal air-coal ratio is determined in conjunction with thermal efficiency, and early warnings are generated and incorporated into the collaborative optimization engine's optimization strategy. The predictive maintenance model is used to generate specific maintenance suggestions based on the aligned boiler operating status monitoring data received by the multi-source data fusion module and output them to the collaborative optimization engine. The collaborative optimization engine integrates information generated by the AI ​​visual recognition algorithm module, the carbon emission-combustion linkage algorithm module, the predictive maintenance model, and the multi-dimensional linkage algorithm module to generate optimization strategies.

[0027] See Figure 4 The present invention also provides a monitoring and control method based on the above-mentioned intelligent monitoring system for thermal power boilers, comprising the following steps: S1: The boiler intelligent monitoring terminal collects boiler operation status monitoring data, including multi-dimensional physical parameters and AI visual image data inside the furnace. The monitoring data is then analyzed locally and executed control commands are sent to the multi-dimensional associated control device. At the same time, the monitoring data is uploaded to the cloud collaborative analysis platform. S2: Collect data by linking the multi-dimensional correlation control device with the intelligent boiler monitoring terminal, execute the execution control command, and generate early warning information and graded response command based on the analysis results of the intelligent boiler monitoring terminal or the cloud collaborative analysis platform, and send them to the graded response terminal; S3: Receive the warning information and graded response instructions using the graded response terminal, and execute the corresponding graded response operation; S4: Utilize the cloud-based collaborative analysis platform to analyze and process the received monitoring data, generate a collaborative optimization strategy encompassing four dimensions: safety, energy efficiency, environmental protection, and maintenance, and distribute the optimization strategy to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control device to iteratively update the control parameters.

[0028] See Figure 5 Taking a boiler monitoring system as an example, in the implementation of the case, the boiler intelligent monitoring terminal integrates a multi-dimensional parameter acquisition module and an AI visual monitoring module. The parameter acquisition module is equipped with a high-precision thermocouple sensor, pressure transmitter, gas sensor, liquid level sensor, ultrasonic vibration sensor, and vibration sensor. The AI ​​visual monitoring module uses a high-temperature resistant infrared high-definition camera (temperature resistance ≥1200℃) and a lidar to collect image data of furnace flame morphology, furnace wall coking status, and tube screen surface defects.

[0029] Among them, the parameter acquisition module, through the coordinated deployment of multiple types of sensors, can capture core parameters such as temperature field distribution, medium pressure, flue gas composition, medium liquid level and equipment vibration amplitude in real time during boiler operation. The sampling frequency is up to 1 second / time, and the data error is controlled within ±0.5%, providing quantitative basic data for boiler operation status. The AI ​​visual monitoring module, through the combination of high-temperature resistant infrared camera (adapted to the high-temperature environment of the furnace) and lidar (penetrating high dust interference), accurately collects information on the combustion uniformity of the furnace flame, the area / thickness of coking on the furnace wall, and hidden defects such as microcracks or corrosion on the surface of the tube screen. By locally integrating the quantitative data from parameter acquisition with the image information from visual monitoring, the terminal can achieve dual-dimensional state perception, both numerical and visual. This compensates for the inability of single-parameter monitoring to identify early coking and micro-defects in the tube screen, providing comprehensive raw data support for the generation of subsequent graded early warning and coordinated control instructions.

[0030] The local collaborative analysis unit of the boiler intelligent monitoring terminal has a built-in lightweight fusion model, and sets up a three-level early warning threshold including regular deviation, suspicious anomaly and emergency fault, as well as a dual-condition triggering mechanism for parameter anomaly and visual feature anomaly, and has the ability to store 168 hours of full data.

[0031] The lightweight fusion model of the local collaborative analysis unit adopts a lightweight CNN-LSTM architecture, which can complete the correlation analysis of single-round parameters and visual data within 500ms. The three-level early warning threshold corresponds to differentiated response standards, and adopts a triggering mechanism that triggers local prompt pop-ups on the terminal only for regular deviations, triggers audible and visual alarms on the operation and maintenance terminal for suspicious anomalies, and directly links the power plant DCS system for strong reminders for emergency faults. The dual-condition triggering mechanism of parameter anomalies and visual feature anomalies requires that the quantified parameters exceed the threshold and the AI ​​visually recognizes the corresponding fault features before triggering an early warning, which can effectively reduce the false alarm rate of traditional single-parameter early warning. The 168-hour full data storage retains continuous operating data for 3 days before the fault occurred, providing a complete data chain for fault root cause tracing. Through lightweight fusion computing, hierarchical threshold determination, and dual-condition triggering logic of the local collaborative analysis unit, this terminal can complete a local closed loop of real-time perception, rapid analysis, and accurate early warning without cloud support. This avoids the latency risk of cloud transmission, maintains basic monitoring functions when the network is interrupted, and provides reliable data support for subsequent fault review and model iteration.

[0032] The real-time early warning trigger module of the boiler intelligent monitoring terminal adopts a hierarchical early warning mechanism. Regular deviations are indicated by platform text prompts and parameter adjustment suggestions, suspicious anomalies are indicated by audible and visual prompts and pre-push combustion optimization instructions, and emergency faults are activated by strong light, voice alarms, cloud-based emergency early warnings, and associated equipment protection modes.

[0033] Among them, the platform text prompts for regular deviations will accurately mark the out-of-tolerance parameters, and the synchronously pushed parameter adjustment suggestions will match the boiler's historical best operating database; the audible and visual prompts for suspicious anomalies use a combination of high-frequency buzzer (2 seconds / time) and yellow flashing indicator lights, and the combustion optimization command pre-push will generate 2-3 sets of air supply / coal feeding ratio schemes adapted to the current load, which can be quickly selected by operation and maintenance personnel within 5 minutes; the strong light prompt for emergency faults is a full-screen red flashing on the terminal, the voice alarm will cycle through the fault type, the cloud emergency warning will simultaneously trigger the power plant DCS system pop-up window and the operation and maintenance team's mobile phone alarm, and the associated equipment protection mode will automatically start the protective action; This tiered and differentiated early warning and response design effectively addresses both low-risk anomalies that excessively consume operational resources and high-risk faults that receive multi-dimensional and high-priority handling support. It achieves precise guidance for minor anomalies and rapid linkage for major faults, thereby effectively shortening the average response time for fault handling and reducing the probability of secondary risks caused by misoperation.

[0034] Multi-dimensional correlation control equipment and boiler intelligent monitoring terminal achieve data interoperability through 5G, industrial Ethernet and wireless dual-mode communication, and carbon emission online monitoring instrument collects CO2, SO2 and NO in flue gas. x Concentration data, combustion optimization controller adjusts damper opening and fuel supply, waste heat recovery intelligent regulator dynamically adjusts recovery efficiency, and boiler life cycle archive stores full-cycle data for multi-dimensional comparison.

[0035] The 5G link handles high-bandwidth transmission of 1080P high-definition images and LiDAR point cloud data from the AI ​​visual monitoring module, with transmission latency controlled within 20ms. The industrial Ethernet handles low-latency interaction of quantitative data such as temperature and pressure from the parameter acquisition module, with a data jitter rate of ≤1%. Wireless communication serves as a backup link, automatically switching within 10 seconds in case of a primary link interruption to ensure uninterrupted basic operational data. The carbon emission online monitoring instrument has a sampling frequency of 2 seconds / time, with CO2 concentration monitoring accuracy reaching ±10ppm and SO2 / NO... x With an accuracy of ±5ppm, it can pinpoint pollutant emission peaks in real time; the combustion optimization controller dynamically adjusts the damper opening (adjustment accuracy ±1%) and fuel supply (adjustment step size 0.5t / h) based on the flame combustion uniformity and flue gas composition data transmitted from the monitoring terminal, balancing combustion efficiency and carbon emission levels; the waste heat recovery intelligent regulator dynamically adjusts the heat exchanger heat exchange area ratio based on flue gas outlet temperature and steam pressure, stabilizing waste heat recovery efficiency at over 85%; the boiler's full life cycle archive also synchronously stores equipment factory parameters, operation and maintenance records, fault handling plans, and other information, which can be compared with the best operating data of the same model of boiler and industry benchmark data to pinpoint operational energy efficiency shortcomings; Through dual-mode communication, multi-dimensional interconnected control devices and monitoring terminals can communicate in real time. Each device can respond in a coordinated manner based on the fusion analysis results of the terminal. For example, the excessive data of the carbon emission monitor will trigger the precise adjustment of the combustion optimization controller, and the waste heat recovery efficiency data will be synchronized to the archive for long-term performance evaluation, forming a full-link collaboration to help the boiler reduce carbon emission intensity and improve waste heat recovery efficiency.

[0036] The tiered response system employs a four-level response mechanism. The operation and maintenance management platform and maintenance personnel's mobile phones are the first priority (0-45 seconds response), the power plant's DCS system is the second priority (45-90 seconds response), the environmental monitoring platform is the third priority (90-120 seconds response), and the emergency response center is the fourth priority (responding after 120 seconds). Specifically, the first priority (0-45 seconds) operation and maintenance management platform will display a summary card showing the fault type, location of associated equipment, and real-time operating data. Maintenance personnel's mobile phones will simultaneously receive push notifications containing fault characteristics and preliminary handling suggestions, supporting direct access to 168 hours of historical data stored on the mobile terminal for rapid comparison. If no handling feedback is received from the operation and maintenance end within 45 seconds, the response automatically escalates to the second priority (45-90 seconds). The power plant's DCS system will automatically retrieve the associated parameter curves of the faulty equipment, triggering load limits (fluctuation range ≤5%) and equipment operation protection modes in the corresponding area, and highlighting the fault link on the DCS main interface. If the fault is not resolved within 90 seconds, the response escalates to the third priority. At the first level (90-120 seconds), the environmental monitoring platform will simultaneously receive carbon emissions and pollutant emissions data during the fault period, automatically generate a temporary report on the fault and emissions, and indicate whether there is a risk of exceeding standards. If the fault continues after 120 seconds, the fourth-priority emergency response center will receive complete fault source data (including visual monitoring images and parameter change curves), and simultaneously activate the emergency plan for dispatching on-site repair teams and delineating safety protection areas. This mechanism ensures rapid intervention from the operation and maintenance end, avoids fault response delays through automatic upgrade mechanisms, and links the entire process of production, environmental protection, and emergency response, reducing the connection time from fault triggering to multi-entity collaborative handling to within 2 minutes, effectively reducing the risk of fault escalation.

[0037] The cloud-based collaborative analysis platform's multi-source data fusion module performs standardized cleaning, time-series alignment, and abnormal data filtering on parameter data, visual data, and carbon emission data before synchronously transmitting them to the core algorithm module.

[0038] The multi-source data fusion module sets differentiated processing rules for different data types: For parameter data (temperature, pressure, etc.), the output units of each sensor are uniformly converted to industry standards, and the data accuracy is calibrated to two decimal places; for visual data (furnace flame frames, furnace wall coking images), the image acquisition timestamp is extracted and time-series aligned with the sampling time of the parameter data, with the alignment error controlled within 0.1 seconds to ensure consistency between the time dimension of numerical and visual data; for carbon emission data, outliers with instantaneous fluctuations exceeding 20% ​​are filtered using the 3σ principle. Filtered data is marked for manual verification and stored separately to avoid interfering with normal analysis. After processing, all data is uniformly packaged into a JSON structure with data quality tags (high quality / qualified / to be verified) and synchronously transmitted to the core algorithm module. Through the standardization, alignment, and filtering of the multi-source data fusion module, the originally heterogeneous parameter, visual, and carbon emission data have a unified analytical benchmark. This can effectively reduce the preprocessing computation of the core algorithm module and avoid the interference of abnormal data on the analysis results, providing high-quality unified data input for subsequent core analyses such as combustion optimization prediction and coking trend prediction of the CNN-LSTM model.

[0039] The core algorithm modules of the cloud-based collaborative analysis platform include an AI visual recognition algorithm module that integrates CNN convolutional neural networks, a carbon emission-combustion linkage algorithm module, a predictive maintenance model that integrates LSTM long short-term memory networks, and a multi-dimensional linkage algorithm module that integrates recurrent neural networks. These modules respectively enable functions such as furnace anomaly identification, low-carbon combustion optimization, early fault prediction, and dynamic adjustment of early warning thresholds.

[0040] The AI ​​visual recognition algorithm module, which integrates CNN convolutional neural networks, extracts features (such as flame brightness uniformity and texture / thickness features of the coking area) from images of furnace flames, furnace wall coking, and tube screen defects collected by the visual monitoring module. It enhances feature recognition through a 5-layer convolutional co-pooling layer structure, improving the accuracy of furnace anomaly identification. It can also distinguish between the initial (thickness < 5mm) and middle (5-15mm) stages of coking, providing precise data for handling. The carbon emission-combustion linkage algorithm module combines carbon emission data with combustion parameters (damper opening, fuel supply) to construct a dual-objective optimization model for carbon emission concentration and combustion efficiency. When the flue gas CO2 concentration exceeds the benchmark value by 5%, it automatically adjusts the excess air coefficient to the range of 1.1-1.2. At the same time, it ensures that the combustion efficiency is not less than 92%, achieving a balance between low carbon emissions and high-efficiency combustion; the predictive maintenance model integrating LSTM long short-term memory network, based on more than one year of historical operating parameters (temperature, vibration, pressure) and fault records, learns the equipment deterioration time trend, and can predict faults such as tube screen corrosion and fan bearing wear 7-15 days in advance, and simultaneously outputs targeted maintenance suggestions (such as key inspection areas of tube screen); the multi-dimensional linkage algorithm module integrating recurrent neural network will dynamically update the three-level early warning thresholds by combining factors such as boiler real-time operating load, ambient humidity, and equipment cumulative running time. For example, under high load (≥90% rated power), the furnace temperature early warning threshold can be relaxed by 3%, so that the early warning threshold is adapted to the actual operating scenario and the false alarm rate is reduced; Through the collaborative operation of these four core algorithm modules, the cloud-based collaborative analysis platform can achieve full-link intelligent analysis, including accurate identification of furnace anomalies, dynamic optimization of combustion and carbon emissions, early prediction of faults, and adaptation and adjustment of early warning thresholds. It not only provides strategy support for local terminals but also continuously optimizes algorithm effects through long-term data iteration, thereby helping to improve the overall operating efficiency of boilers.

[0041] The cloud-based collaborative analysis platform's collaborative optimization engine generates a comprehensive collaborative optimization strategy report that integrates safety, energy efficiency, environmental protection, and maintenance. This report is then distributed to terminals and related equipment, while simultaneously updating the boiler's full lifecycle archive to enable dynamic strategy iteration.

[0042] The collaborative optimization strategy report includes four dimensions with clearly defined implementation details: The safety dimension identifies high-risk areas (such as furnace wall sections in the initial coking stage) using AI vision and provides suggestions for temporary operating load limits and equipment protection mode activation; the energy efficiency dimension refines the optimal parameter combinations for damper opening and fuel supply based on carbon emission-combustion linkage algorithm results, and marks the target range for waste heat recovery efficiency under current operating conditions; the environmental protection dimension combines pollutant monitoring data to clarify the adjustment range of denitrification agent injection and the optimized value of flue gas recirculation, ensuring that CO2, SO2, and NO... x Emission concentrations are kept within 80% of the national standard limits; based on predictive maintenance model warnings, maintenance plans are formulated for the near future (such as checking the micro-crack area of ​​the tube panel within 3 working days), and vulnerable parts that need to be replaced first (such as fan bearings with excessive vibration) are marked. The update of the boiler's full life cycle archive will simultaneously record the content of this strategy, changes in operating data after execution, and the effect of fault handling, forming a closed-loop data chain of strategy generation, execution and feedback, accumulating a basis for subsequent strategy iterations; By integrating and distributing multi-dimensional strategies through the collaborative optimization engine, terminals and related devices can synchronously perform matching adjustments (such as combustion optimization controller response damper parameters and waste heat recovery regulator adaptation efficiency targets). At the same time, with the help of dynamic updates to the archive, the engine can continuously optimize the accuracy of strategies based on historical feedback, thereby improving the adaptability of subsequent strategies to boiler operating conditions by 15% and gradually achieving long-term optimal balance in operation.

[0043] The intelligent boiler monitoring terminal and the cloud-based collaborative analysis platform use the national cryptographic SM4 encryption communication protocol (combined with the AES-256 encryption algorithm), along with two-way device identity authentication, dynamic key generation, and data desensitization processing, to ensure the security and privacy of data transmission. Among them, the national cryptographic SM4 encryption communication protocol is mainly used to encrypt the communication link between the terminal and the cloud, ensuring that the data transmission process is not illegally eavesdropped or tampered with, while the AES-256 encryption algorithm performs secondary encryption on sensitive data such as visual images and core operating parameters, forming a double encryption protection; the two-way authentication of device identity requires the terminal and the cloud to verify each other's digital certificates (the certificates are issued by the power plant's internal CA center and are valid for 90 days). The authentication process includes three random code verifications. If either party fails to authenticate, the communication link will be immediately disconnected and a security alarm will be triggered; dynamic key generation uses the device's unique hardware identifier and the current timestamp as seeds, and updates the session key every 10 minutes to avoid the risk of leakage caused by long-term use of fixed keys; data anonymization processing will anonymize and replace private content such as the specific name of the power plant and the information of operation and maintenance personnel in the transmitted data, retaining only the necessary business information such as the device number and operating parameters, while masking the three redundant decimal places in the parameter data, reducing the risk of privacy leakage while ensuring the usability of analysis; Through the aforementioned multi-layered security mechanisms, the data interaction between the terminal and the cloud not only meets the Level 3 requirements of the Cybersecurity Classified Protection 2.0, but also effectively avoids risks such as unauthorized access, data eavesdropping, and privacy leaks, minimizing the incidence of data security incidents without affecting the efficiency of data transmission.

[0044] When implementing this procedure, please follow these steps: First, deploy and install the intelligent boiler monitoring terminal and multi-dimensional associated control equipment, and complete the hardware networking and communication debugging; Then, deploy and install the boiler intelligent monitoring terminal and multi-dimensional associated control equipment, and complete the hardware networking and communication debugging; Reactivate the tiered response terminal, configure early warning rules and response procedures, and start the system for real-time monitoring and analysis; Finally, the system was run for joint debugging and testing to verify the full-link functions of data acquisition, early warning triggering, linkage control and policy distribution.

[0045] In summary, this invention provides an intelligent monitoring system and method for thermal power boilers. By integrating multi-dimensional parameter acquisition and AI visual monitoring through a boiler intelligent monitoring terminal, it achieves real-time intelligent identification of furnace flame morphology, coking status, and tube panel defects. Combined with a local three-level early warning and dual-condition triggering mechanism, it significantly improves the early detection capability of abnormal states. The system uses multi-mode communication such as 5G and industrial Ethernet to link combustion optimization, carbon emission monitoring, waste heat recovery, and other multi-dimensional control equipment, forming a closed loop of data acquisition, analysis, and control. The cloud platform, based on multi-source data fusion and AI algorithms, outputs collaborative optimization strategies including safety, energy efficiency, environmental protection, and maintenance. It also utilizes a tiered response terminal to achieve rapid collaborative handling, effectively improving the safety, stability, and economy of boiler operation, and realizing a shift from passive response to proactive prediction in operation and maintenance. Furthermore, through the carbon emission-combustion linkage algorithm and predictive maintenance model, the system achieves low-carbon optimization and early warning of faults while ensuring the safe operation of the boiler. The hierarchical response mechanism, combined with four-level priority settings, ensures rapid multi-level collaboration from on-site operation and maintenance to environmental supervision, improving emergency response efficiency. The system continuously accumulates data through a full lifecycle archive, supporting dynamic strategy iteration and equipment health management. In addition, it adopts national cryptographic encryption and two-way authentication mechanisms to ensure secure and reliable data communication. Overall, the system realizes intelligent collaborative management and control of thermal power boilers in the dimensions of safety, energy efficiency, environmental protection, and maintenance, providing comprehensive support for the green, efficient, and reliable operation of power plants.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the technical solution of the present invention in any way. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can be modified and replaced in several simple ways, and these modifications and replacements are all within the scope of protection covered by the claims.

Claims

1. An intelligent monitoring system for thermal power boilers, characterized in that, This includes intelligent boiler monitoring terminals, multi-dimensional correlation control equipment, hierarchical response terminals, and cloud-based collaborative analysis platforms; The intelligent boiler monitoring terminal is communicatively connected to the multi-dimensional correlation control device and the cloud-based collaborative analysis platform, respectively. It is used to collect boiler operating status monitoring data and perform local collaborative analysis. Based on the local collaborative analysis results, it sends execution control commands to the multi-dimensional correlation control device and uploads the monitoring data to the cloud-based collaborative analysis platform. The multi-dimensional correlation control device is communicatively connected to the hierarchical response terminal and simultaneously to the boiler intelligent monitoring terminal. It is used to collect data in conjunction with the boiler intelligent monitoring terminal, execute the execution control command, generate early warning information and hierarchical response command based on the analysis results of the boiler intelligent monitoring terminal or the cloud collaborative analysis platform, and send the early warning information and hierarchical response command to the hierarchical response terminal. The tiered response terminal is used to receive the warning information and tiered response instructions, and to execute the corresponding tiered response operations; The cloud-based collaborative analysis platform is used to receive the monitoring data uploaded by the boiler intelligent monitoring terminal, analyze and process the monitoring data to generate optimization strategies, and then distribute the optimization strategies to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control device.

2. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The intelligent boiler monitoring terminal integrates a multi-dimensional parameter acquisition module and an AI visual monitoring module. The multi-dimensional parameter acquisition module is equipped with a high-precision thermocouple sensor, pressure transmitter, gas sensor, liquid level sensor, ultrasonic vibration sensor, and vibration sensor. The AI ​​visual monitoring module uses a high-temperature resistant infrared high-definition camera and lidar with a temperature resistance of not less than 1200℃ to collect image data of furnace flame morphology, furnace wall coking status, and tube screen surface defects.

3. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The intelligent boiler monitoring terminal has a built-in local collaborative analysis unit, which has a built-in lightweight fusion model. The lightweight fusion model has three levels of early warning thresholds: regular deviation, suspicious anomaly, and emergency fault. It adopts a dual-condition triggering mechanism that requires both parameter anomaly and visual feature anomaly to be met simultaneously. The local collaborative analysis unit has a full data storage capacity of at least 168 hours.

4. The intelligent monitoring system for thermal power boilers according to claim 3, characterized in that, The intelligent boiler monitoring terminal also includes a real-time early warning triggering module. The real-time early warning triggering module adopts a hierarchical early warning mechanism, which triggers text prompts and parameter adjustment suggestions on the platform for regular deviations, triggers audible and visual prompts and pre-pushing of combustion optimization instructions for suspicious abnormalities, and triggers strong light and voice alarms, cloud-based emergency early warnings and activation of related equipment protection modes for emergency faults.

5. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The multi-dimensional interconnected control equipment includes a combustion optimization controller, an online carbon emission monitor, a waste heat recovery intelligent regulator, a furnace tube corrosion monitoring module, a water supply system controller, an induced draft / forced draft regulator, a blowdown valve control module, and a boiler lifecycle archive. The combustion optimization controller, carbon emission online monitor, waste heat recovery intelligent regulator, furnace tube corrosion monitoring module, water supply system controller, induced draft / forced draft regulator, and blowdown valve control module are all communicatively connected to the boiler intelligent monitoring terminal, and are linked with the boiler intelligent monitoring terminal to collect data and execute the execution control commands. The boiler lifecycle archive is used to store equipment lifecycle data for multi-dimensional comparison.

6. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The multi-dimensional correlation control device and the boiler intelligent monitoring terminal achieve data interconnection through dual-mode communication of 5G, industrial Ethernet and wireless communication.

7. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The tiered response terminals include an operation and maintenance management platform, mobile phones for operation and maintenance personnel, a power plant DCS system, an environmental protection supervision platform, and an emergency response center.

8. The intelligent monitoring system for thermal power boilers according to claim 7, characterized in that, The tiered response terminal adopts a four-level response plus automatic upgrade mechanism. The operation and maintenance management platform and the mobile phones of operation and maintenance personnel are configured as the first priority, with a response time of 0 to 45 seconds; the power plant DCS system is configured as the second priority, with a response time of 45 to 90 seconds; the environmental protection supervision platform is configured as the third priority, with a response time of 90 to 120 seconds; and the emergency response center is configured as the fourth priority, with a response time after 120 seconds.

9. The intelligent monitoring system for thermal power boilers according to claim 1, characterized in that, The cloud-based collaborative analysis platform includes a multi-source data fusion module, an AI visual recognition algorithm module, a carbon emission-combustion linkage algorithm module, a predictive maintenance model, a multi-dimensional linkage algorithm module, and a collaborative optimization engine. The multi-source data fusion module is used to align, clean, and distribute the acquired boiler operation status monitoring data to downstream modules; The AI ​​visual recognition algorithm module is used to receive aligned boiler operation status monitoring data from the multi-source data fusion module for identification, and generate a structured report containing fault type, location, confidence level, and timestamp for collaborative optimization engine decision-making; The carbon emission-combustion linkage algorithm module is used to construct a mapping relationship between fuel quantity, air volume, load, and emission concentration based on the aligned boiler operation status monitoring data received from the multi-source data fusion module. It then calculates in real time the solution to minimize CO2 emissions or reduce NO emissions under the current load. x The optimal air-coal ratio with the best overall thermal efficiency is determined, and early warnings are generated and incorporated into the collaborative optimization engine's optimization strategy. The predictive maintenance model is used to generate specific maintenance suggestions based on the aligned boiler operating status monitoring data received by the multi-source data fusion module and output them to the collaborative optimization engine. The collaborative optimization engine integrates information generated by the AI ​​visual recognition algorithm module, the carbon emission-combustion linkage algorithm module, the predictive maintenance model, and the multi-dimensional linkage algorithm module to generate optimization strategies.

10. A monitoring and control method based on the intelligent monitoring system for thermal power boilers according to any one of claims 1-9, comprising the following steps: Boiler intelligent monitoring terminal collects boiler operation status monitoring data, including multi-dimensional physical parameters and AI visual image data inside the furnace. The monitoring data is then analyzed locally and executed control commands are sent to the multi-dimensional associated control device. At the same time, the monitoring data is uploaded to the cloud collaborative analysis platform. The system utilizes multi-dimensional correlation control devices to collect data in conjunction with the intelligent boiler monitoring terminal, executes the execution control commands, and generates early warning information and graded response commands based on the analysis results of the intelligent boiler monitoring terminal or the cloud-based collaborative analysis platform, which are then sent to the graded response terminal. The system utilizes a tiered response terminal to receive the warning information and tiered response instructions, and executes the corresponding tiered response operations. The received monitoring data is analyzed and processed using a cloud-based collaborative analysis platform to generate a collaborative optimization strategy encompassing four dimensions: safety, energy efficiency, environmental protection, and maintenance. This optimization strategy is then distributed to the boiler intelligent monitoring terminal and / or the multi-dimensional associated control equipment to iteratively update the control parameters.

Images