Thermal power plant coal burning whole process management system based on AI intelligent identification

Abstract

The invention discloses a thermal power plant coal burning whole process management system based on AI intelligent identification, and the system comprises a sampling room AI intelligent identification system which identifies the number of people entering a sampling room, and judges whether to carry out the sampling of two persons; the sample preparation room AI intelligent identification system is used for identifying the number of personnel entering the sample preparation room, judging whether double-person sample preparation is carried out, judging whether the coal sample drying time is compliant, and analyzing and judging whether coal sample weighing data is compliant; the AI intelligent identification system for full-moisture sample preparation identifies whether a riffle is adopted for division or not in the sample preparation process by using an AI technology; and the AI intelligent identification system of the calorific value preparation chamber identifies whether an operator has the action of checking the oxygen bomb and the crucible or not when the oxygen bomb is opened after the test is finished. According to the invention, the AI intelligent identification technology is introduced into each key link of the fire coal from purchasing to burning into ash, so that automatic monitoring of sampling, sample preparation, testing and other processes is realized, the operation is ensured to accord with specifications, and the accuracy and safety of fire coal management are improved.

Description

Technical Field

[0001] This invention relates to the field of coal-fired power plant management technology, specifically to a coal-fired power plant whole-process management system based on AI intelligent recognition. Background Technology

[0002] In the operation of thermal power plants, coal management is crucial. Traditional coal management processes rely heavily on manual operations, which presents numerous problems. For example, in the sampling stage, it is difficult to ensure the strict implementation of dual-person sampling, leading to potential deviations in sampling results; during sample preparation, the compliance of coal sample drying time and weighing data lacks effective real-time monitoring; whether a divider is used for reduction during total moisture sampling is also difficult to control precisely; in the calorific value preparation room, after the experiment, it is impossible to promptly and effectively supervise whether the operators check the oxygen bomb and crucible to determine whether the coal sample has burned completely. These problems can affect the accuracy of coal quality testing, thereby affecting the power plant's combustion efficiency and economic benefits, and may even pose safety hazards. With the development of technology, AI intelligent recognition technology has gradually matured, providing new approaches to solving these problems. Summary of the Invention

[0003] To address these issues, this invention provides an AI-based intelligent recognition-based management system for the entire coal combustion process in thermal power plants.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] The AI-based intelligent recognition-based coal combustion management system for thermal power plants includes:

[0006] The sampling room AI intelligent recognition system uses high-definition cameras and AI image recognition equipment to identify the number of people entering the sampling room, determine whether two people are sampling, and monitor the sampling operation process;

[0007] The sample preparation room AI intelligent recognition system uses high-definition cameras and AI image recognition equipment to identify the number of people entering the sample preparation room, determine whether two people are preparing samples, and at the same time, through monitoring the operating status of the sample preparation equipment and image recognition, determine whether the coal sample drying time is compliant, and analyze the coal sample weighing data to determine whether it is compliant.

[0008] The AI ​​intelligent recognition system for total moisture sample preparation installs an image recognition device at the total moisture sample preparation equipment and uses AI technology to identify whether a divider is used for sample reduction during the sample preparation process.

[0009] The AI ​​intelligent recognition system in the calorific value preparation room uses cameras and AI intelligent analysis equipment to identify whether the operator has checked the oxygen bomb and crucible when the oxygen bomb is opened after the test, and to determine whether the coal sample has been completely burned.

[0010] Each batch of coal is assigned a unique identifier from the moment it is sampled upon arrival at the plant. Data related to each batch of coal is recorded sequentially by the AI ​​intelligent recognition systems in the sampling room, sample preparation room, total moisture sample preparation room, and calorific value preparation room. Simultaneously, the monitoring records are linked together to form a complete and tamper-proof digital quality history.

[0011] Furthermore, the AI ​​image recognition device of the sampling room AI intelligent recognition system is pre-trained with personnel recognition and operation process recognition models, which are used to analyze and judge the personnel entering the sampling room and the sampling operation.

[0012] Furthermore, the AI ​​intelligent recognition system in the sample preparation room installs temperature and humidity sensors on the drying equipment, inputs environmental parameters and coal sample image data collected by the camera into the AI ​​algorithm, and determines whether the coal sample drying time is compliant.

[0013] Furthermore, the image acquisition device of the AI ​​intelligent recognition system for total moisture sample preparation is installed at the reduction section of the total moisture sample preparation equipment to clearly capture the reduction operation process for analysis by the AI ​​image recognition system.

[0014] Furthermore, the AI ​​algorithm of the AI ​​intelligent recognition system in the calorific value preparation room analyzes the operator's hand movements and body posture to determine whether there are actions to check the oxygen bomb and crucible, and analyzes the images of the oxygen bomb and crucible to determine the combustion state of the coal sample.

[0015] The present invention has the following advantages: By introducing AI intelligent recognition technology into each key link of coal from procurement to combustion into ash, the present invention realizes automated monitoring of sampling, sample preparation, and testing processes, ensuring that the operation complies with the specifications and improving the accuracy and safety of coal management.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. Attached Figure Description

[0017] To more intuitively illustrate the prior art and this application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing this application; for example, based on the technical concept disclosed in this application and the exemplary drawings, those skilled in the art are able to easily make conventional adjustments or further optimizations to the addition / reduction / classification, specific shapes, positional relationships, connection methods, size ratios, etc. of certain units (components).

[0018] Figure 1This is a system block diagram of an AI-based intelligent recognition-based coal-fired power plant whole-process management system provided in one embodiment of this application. Detailed Implementation

[0019] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these embodiments are merely for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Technical engineers in the field can make some non-essential improvements and adjustments to the present invention based on the above-described content. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 The AI-based intelligent recognition system for the entire coal combustion process in thermal power plants includes an AI intelligent recognition system for the sampling room, an AI intelligent recognition system for the sample preparation room, an AI intelligent recognition system for the total moisture sample preparation room, and an AI intelligent recognition system for the calorific value preparation room.

[0021] The sampling room's AI-powered intelligent recognition system installs high-definition cameras at the entrance and key locations inside the sampling room to ensure clear capture of personnel entering and the sampling process. The image data captured by the cameras is transmitted to an AI image recognition device, which is pre-trained with models for personnel and operational process recognition. When personnel enter the sampling room, the model analyzes the image to determine if there are two people present.

[0022] Meanwhile, during the sampling process, the model monitors the actions of the sampling personnel in real time and compares them with the preset standard operating procedures. If any violation is found, an alarm message is immediately sent to the management system, and the time and specific circumstances of the violation are recorded.

[0023] The sampling room AI intelligent recognition system employs multimodal fusion recognition technology, integrating visible light imaging, infrared thermal imaging, and millimeter-wave radar to construct a three-dimensional spatial personnel positioning model. It uses the YOLOv8 target detection algorithm for personnel counting and identity verification, and combines a pose estimation network (an improved version of OpenPose) to analyze the standardization of sampling actions. The system is equipped with an edge computing unit, enabling real-time inference to be performed locally.

[0024] In addition, the AI ​​intelligent recognition system in the sampling room can automatically adjust the image enhancement parameters according to the ambient light intensity to solve the problem of false detection caused by insufficient light in the sampling room; it establishes a standard action sequence template for sampling operations, compares the real-time operation flow through an LSTM network, and identifies missed steps / out-of-order operations; it adopts an attention mechanism to enhance the extraction of interaction features between two people, distinguish the roles of the main and auxiliary operators, and prevent substitution and cheating.

[0025] The formula for calculating the sampling compliance judgment index is as follows:

[0026] ;

[0027] Where, N detected N represents the actual number of valid operation steps detected. required W is the total number of steps specified in the standard process. pose K represents the posture correctness weighting coefficient (0~1). sequence This is a correction factor for the order of operations.

[0028] AI intelligent recognition system for sample preparation room: Install cameras and AI image recognition equipment similar to those in the sampling room in the sample preparation room.

[0029] To determine the drying time of coal samples, environmental parameters during the drying process are acquired by installing temperature and humidity sensors on the drying equipment. These parameters, along with coal sample image data captured by a camera, are then input into an AI algorithm. Based on a preset drying model and empirical data, the AI ​​algorithm analyzes the drying state of the coal sample and determines whether the drying time is compliant.

[0030] When weighing coal samples, the weighing sensor is connected to the AI ​​analysis device, and the weighing data is transmitted to the AI ​​device in real time. The AI ​​device determines whether the weighing data is compliant based on the preset compliance data range and allowable fluctuation value.

[0031] The sample preparation room AI intelligent recognition system constructs a sample preparation process simulation platform based on digital twins, integrating PLC controller data and visual perception information; it uses the Transformer architecture to process spatiotemporal sequence data, realizing precise control of the coal sample drying process.

[0032] The sample preparation room's AI intelligent recognition system can achieve multi-source heterogeneous data fusion, transfer learning optimization, and weighing anomaly detection.

[0033] Among them, multi-source heterogeneous data fusion, simultaneous acquisition of temperature (±0.5℃ accuracy), humidity (±2%RH accuracy), and image features (RGB+HSV color space) to construct a three-dimensional dry state characterization vector;

[0034] Transfer learning optimization utilizes pre-trained ResNet-50 models for domain adaptation, and establishes a differentiated drying model library for lignite / bituminous coal / anthracite.

[0035] For weighing anomaly detection, the Isolation Forest algorithm is introduced to monitor changes in the entropy value of the weighing curve in real time and identify sudden quality fluctuations.

[0036] The dryness prediction model is as follows:

[0037] ;

[0038] Where D(t) is the predicted dryness of the coal sample at time t; A is the initial moisture content influencing factor; B is the environmental temperature and humidity coupling coefficient; and C is the material property correction term.

[0039] The formula for assessing the reliability of weighing data is:

[0040] ;

[0041] Where σ is the standard deviation of the weighing data; μ is the theoretical mean; Δt is the weighing time deviation; and τ is the time constant.

[0042] Total Moisture Sample Preparation AI Intelligent Recognition System: A dedicated image acquisition device is installed at the reduction section of the total moisture sample preparation equipment. This device can clearly capture the reduction operation process. The AI ​​image recognition system analyzes the acquired images in real time to identify whether the equipment is a divider and whether the divider operation is correct. If it is found that the divider is not being used or the operation is not standardized, an alarm is immediately issued to the operator, and the information is fed back to the management system for timely correction.

[0043] The AI ​​intelligent recognition system for full-moisture sample preparation uses instance segmentation technology (Mask R-CNN) to achieve pixel-level analysis of the binary divider operation process; it captures the material movement trajectory through a high-speed camera (500fps) and verifies the uniformity of the reduction by combining fluid dynamics simulation.

[0044] The core functional modules of the total moisture sample preparation AI intelligent recognition system include divider morphology recognition, material distribution balance calculation, and anti-blocking early warning mechanism.

[0045] Divider shape recognition: Establish a CAD model library of different types of dividers, and train a point cloud registration network using the ShapeNet dataset;

[0046] Material distribution balance calculation: Based on fractal dimension theory, the material distribution characteristics are quantified, requiring the mass difference between the left and right sides to be ≤0.5%;

[0047] The formula for calculating the uniformity index of the reduction is:

[0048] ;

[0049] Where U represents the uniformity of single-step reduction; For the collection quality of each receiving hopper; denoted as average mass; n represents the total number of receiving hoppers.

[0050] Anti-blockage early warning mechanism: Monitor the equipment's operating status through vibration sensor spectrum analysis (FFT transform) and provide an early warning of blockage risk 30 seconds in advance.

[0051] The formula for predicting the probability of congestion is:

[0052] ;

[0053] Among them, E kin E represents the kinetic energy of the material. crit θ represents the critical unblocking energy; θ is the deflector tilt angle function.

[0054] AI Intelligent Recognition System for Calorific Value Preparation Room: Cameras and AI intelligent analysis equipment are installed in the calorific value preparation room. The AI ​​intelligent analysis equipment processes the images captured by the cameras in real time. When the operator opens the oxygen bomb, the AI ​​algorithm analyzes the operator's hand movements and body posture to determine whether there is any action to check the oxygen bomb and crucible. Furthermore, it analyzes the images of the oxygen bomb and crucible to determine whether the coal sample has burned completely. If any failure to check or incomplete combustion of the coal sample is detected, the system promptly alerts the operator and feeds the relevant information back to the management system.

[0055] The AI-powered intelligent recognition system in the calorimetric preparation chamber employs multispectral imaging technology (covering visible light, near-infrared, and short-wave infrared) to construct a perspective model of the internal combustion state of the oxygen bomb. It extracts the temperature field distribution characteristics of the crucible wall using a convolutional neural network and combines this with acoustic signature analysis to determine the intensity of combustion.

[0056] The AI ​​intelligent recognition system in the calorific value preparation room can determine combustion safety, detect the oxygen bomb's sealing performance, and break down operational actions.

[0057] Among them, the determination of combustion completeness is based on three criteria: the combined use of residual char area ratio (threshold > 95%), NOx generation inversion, and flame duration.

[0058] The formula for calculating complete combustion rate is:

[0059] ;

[0060] Among them, A residue A represents the projected area of ​​the residue. total denoted as the crucible bottom area; k is the temperature decay coefficient; ΔT is the temperature drop rate after extinguishing the flame.

[0061] Oxygen bomb sealing test uses a pressure sensor to plot the charge / discharge PV curve and locates the leak location by the point of abrupt change in slope;

[0062] The formula for estimating oxygen bomb leakage is:

[0063] ;

[0064] Where Δm is the mass of gas leaked; ρ is the oxygen density; V is the volume of the gas chamber; and τ is the time constant.

[0065] Operation decomposition is based on the Spatiotemporal Action Detection Network (STASSN) to break down the opening operation into 7 atomic actions and verify the operation's compliance frame by frame.

[0066] This invention can improve the accuracy of coal management. Through AI intelligent recognition technology, it can accurately determine whether the operations of sampling, sample preparation, and testing meet the standards, reduce errors caused by human factors, improve the accuracy of coal quality testing, and thus provide a more reliable basis for power plant combustion, improve combustion efficiency, and reduce energy consumption.

[0067] Enhance safety; enable timely detection of violations during operation, such as single-person sampling or failure to inspect oxygen bombs, to avoid safety accidents caused by improper operation and protect the personal safety of power plant personnel and equipment safety;

[0068] Improve management efficiency; automated monitoring systems can record and provide feedback on information from each stage in real time, eliminating the need for frequent manual checks and recordings, thus greatly improving the efficiency of coal management and saving labor costs.

[0069] In this invention, the sample preparation room AI intelligent recognition system receives the original coal sample from the sampling room and ensures that a general-purpose analytical coal sample that can be used for various analyses (such as total moisture, calorific value, ash content, etc.) is prepared by monitoring processes such as drying, crushing, and mixing; the output coal sample is the "master sample" for multiple subsequent analytical projects.

[0070] The AI-powered intelligent identification system for total moisture sampling extracts a portion of the coal sample early in the sampling process, specifically for determining total moisture content. The system focuses on monitoring the most critical "reduction" operation to ensure the representativeness of the moisture sample is not compromised.

[0071] Calorific value is one of the core indicators for measuring the quality of coal. The AI ​​intelligent recognition system in the calorific value preparation room is located in the final stage of analysis and testing. It examines the final coal sample prepared by the AI ​​intelligent recognition system in the sample preparation room.

[0072] The AI ​​intelligent recognition system in the sampling room, the AI ​​intelligent recognition system in the sample preparation room, the AI ​​intelligent recognition system in the full moisture sample preparation room, and the AI ​​intelligent recognition system in the calorific value preparation room are all integrated into a unified "coal combustion whole process management system" platform, realizing centralized monitoring and collaborative management.

[0073] Each batch of coal is assigned a unique identifier from the moment it is sampled upon entering the plant. The monitoring records of this batch in all four stages (whether it is operated by two people, whether the drying time is compliant, whether a splitter is used, whether the oxygen bomb inspection is in place, etc.) are linked together to form a complete and tamper-proof "digital quality history".

[0074] Collaborative early warning and root cause analysis: When an anomaly is detected in a certain link (such as calorific value testing), the administrator can retrieve the monitoring records and AI analysis results of all upstream links of the batch of coal samples with one click. This greatly simplifies the problem tracing process and can quickly determine whether it is an operational error in the current link or a hidden danger in the upstream link (such as uneven sample preparation leading to poor sample representativeness), realizing the upgrade from "single point alarm" to "full chain root cause analysis".

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A coal-fired power plant full-process management system based on AI intelligent identification, characterized in that, The application relates to a coal quality digital quality management system. The sampling room AI intelligent identification system identifies the number of personnel entering the sampling room through a high-definition camera and an AI image identification device, judges whether double sampling is performed, and monitors the sampling operation process. The sample preparation room AI intelligent identification system identifies the number of personnel entering the sample preparation room through a high-definition camera and an AI image identification device, judges whether double sample preparation is performed, and judges whether the sample drying time is in compliance through sample preparation equipment operation state monitoring and image identification. The full-moisture sample preparation AI intelligent identification system installs an image identification device at a full-moisture sample preparation device, and identifies whether a two-division device is used for division during sample preparation through AI technology. The heat preparation room AI intelligent identification system identifies whether the operator checks the oxygen bomb and the crucible after the oxygen bomb is opened at the end of the test through a camera and an AI intelligent analysis device, and judges whether the coal sample is completely burned. Each batch of coal is assigned a unique identifier from the beginning of sampling in the plant. Each batch of coal-related data is recorded by the sampling room AI intelligent identification system, the sample preparation room AI intelligent identification system, the full-moisture sample preparation AI intelligent identification system, and the heat preparation room AI intelligent identification system in turn. At the same time, the monitoring records are associated to form a complete and tamper-proof digital quality record. 2.The AI intelligent identification-based coal-fired whole-process management system for thermal power plants according to claim 1, characterized in that, The AI image identification device of the sampling room AI intelligent identification system is pre-trained with personnel identification and operation process identification models for analyzing and judging the personnel and sampling operation entering the sampling room. 3.The AI intelligent identification-based coal-fired whole-process management system for thermal power plants according to claim 1, characterized in that, The sample preparation room AI intelligent identification system installs temperature sensors and humidity sensors on the drying equipment, inputs the environmental parameters and the coal sample image data collected by the camera into the AI algorithm, and judges whether the sample drying time is in compliance. 4.The AI intelligent identification-based coal-fired whole-process management system for thermal power plants according to claim 1, characterized in that, The image acquisition device of the full-moisture sample preparation AI intelligent identification system is installed at the division part of the full-moisture sample preparation device, and is used for clearly shooting the division operation process for the AI image identification system to analyze. 5.The AI intelligent identification-based coal-fired whole-process management system for thermal power plants according to claim 1, characterized in that, The AI algorithm of the heat preparation room AI intelligent identification system analyzes the hand movements and body postures of the operator, judges whether the operator checks the oxygen bomb and the crucible, and analyzes the images of the oxygen bomb and the crucible to judge the coal sample burning state.

Images