Coal-fired boiler intelligent optimization method, system and equipment based on multi-source perception and medium

Abstract

The invention discloses a coal-fired boiler intelligent optimization method, system, equipment and medium based on multi-source perception, and belongs to the technical field of thermal power generation, and the method comprises the steps: carrying out the collection of boiler coal data of a coal-fired boiler, and carrying out the feedback compensation; collecting hearth data to reconstruct a combustion field, formulating an optimization control strategy, and constructing an optimization target to perform tail flue multi-component cooperative control; and carrying out dynamic adjustment based on an implementation result of the control strategy, and carrying out real-time monitoring to carry out coking early warning. According to the method, full-closed-loop intelligent control of coal quality perception, combustion optimization and coking prevention and control is formed, and by means of edge calculation layering processing, the boiler efficiency is improved, the power supply coal consumption is reduced and the cost is saved while it is ensured that emission reaches the standard.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal power generation, and in particular to a method, system, equipment and medium for intelligent optimization of a coal-fired boiler based on multi-source perception. Background Art

[0002] Coal-fired boilers are core equipment for thermal power generation, and their combustion efficiency and pollutant emission control have long relied on distributed control systems based on static models. In recent years, with the development of sensing technology, furnace temperature monitoring, offline coal quality analysis, and online tail flue gas detection have been gradually applied, providing a data basis for combustion optimization. However, existing technologies are mostly limited to the processing of a single data source, such as constructing a two-dimensional temperature field model through infrared temperature measurement, or presetting air-coal ratio parameters based on coal quality laboratory sampling results. Although some advanced systems have introduced neural networks to predict coking trends, they are limited by insufficient data fusion depth and delayed control closed-loop response, making it difficult to achieve dynamic optimization of all operating conditions. In addition, tail pollutant control mostly adopts independent closed-loop strategies (such as the separate regulation of SCR denitrification and electrostatic precipitator), lacking a coordinated optimization mechanism for combustion sources and terminal emissions, resulting in conflicting control objectives (such as the contradiction between reducing nitrogen oxides and improving efficiency).

[0003] The shortcomings of existing technologies are concentrated in four areas: First, there is a lack of adaptability to coal quality. Traditional coal quality analysis relies on offline laboratory testing, with update cycles lasting several hours. This fails to respond to real-time fluctuations in the coal entering the furnace, resulting in a mismatch between the preset air-to-coal ratio and actual combustion requirements, causing efficiency losses and pollutant fluctuations. Second, the combustion field is flat. Furnace temperature monitoring often uses two-dimensional scanning, which struggles to capture temperature gradients and localized high-temperature zones within three dimensions. This can easily lead to coking or localized underburning, and it cannot accurately guide burner stratification and leveling. Third, tail flue pollutant control relies on fixed threshold triggering and adjustment, without a coupled model for combustion efficiency, CO generation, and dust emissions. This leads to delayed adjustment and even conflicting objectives. Fourth, there is a disconnect between coking warning and control. Existing coking prediction models are mostly based on historical statistical data and lack dynamic input of real-time three-dimensional temperature fields and coal ash properties. They rely on manual intervention and adjustment after warnings, resulting in poor response timeliness and insufficient control precision. The above-mentioned defects have caused coal-fired boilers to face long-term bottlenecks in efficiency improvement (typical plant electricity consumption rate exceeds 6%) and pressure to meet environmental protection standards (especially frequent instantaneous NOx / CO exceeding the standard under low-load conditions), and there is an urgent need to achieve breakthroughs through multi-source perception fusion and intelligent collaborative control. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the purpose of the present invention is to solve the problems of low combustion efficiency, high pollutant emissions, and equipment coking in coal-fired boilers.

[0006] To solve the above technical problems, the present application provides the following technical solutions: a coal-fired boiler intelligent optimization method based on multi-source perception, comprising,

[0007] The coal-fired boiler coal data is collected and feedback compensation is performed; the hearth data is collected to reconstruct the combustion field, an optimization control strategy is formulated, an optimization target is constructed to perform multi-component collaborative control of the tail flue, dynamic adjustment is performed based on the implementation results of the control strategy, and coking early warning is monitored in real time.

[0008] As a preferred scheme of the coal-fired boiler intelligent optimization method based on multi-source perception, the feedback compensation comprises,

[0009] The collection device is installed to collect the coal data and perform primary detection to input the edge calculation to obtain the coal type characteristic parameters;

[0010] The update time is set to update the coal type characteristic parameters;

[0011] The coal quality compensation is performed by dynamically correcting the wind-coal ratio reference value according to the real-time obtained coal type characteristic parameters.

[0012] As a preferred scheme of the coal-fired boiler intelligent optimization method based on multi-source perception, the hearth data reconstruction comprises,

[0013] The hearth data is collected again by positioning and input to the edge calculation;

[0014] The hearth three-dimensional temperature field model is reconstructed based on the hearth data and the reconstruction time is set;

[0015] Different temperature zones are identified by the three-dimensional temperature field model.

[0016] As a preferred scheme of the coal-fired boiler intelligent optimization method based on multi-source perception, the optimization control strategy comprises,

[0017] According to the three-dimensional temperature field model identification analysis result, if there is a high temperature zone, the edge calculation sends a cooling instruction to perform opening degree precision adjustment;

[0018] If there is a low temperature zone or local uneven burning, the opening degree is increased.

[0019] As a preferred scheme of the coal-fired boiler intelligent optimization method based on multi-source perception, the multi-component collaborative control of the tail flue comprises that the multi-component detection unit is installed at the air preheater outlet of the tail flue, the concentrations of NO x , CO, O2 and dust are monitored in real time by the four-in-one sensor, and the data is fed back to the edge calculation unit in real time.

[0020] When NO x Emission concentration is less than the emission threshold of 50mg / Nm 3 When the burnout wind angle adjustment mechanism is activated, the edge computing unit adjusts the burnout wind angle within the range of ±30 degrees through the burnout wind angle driver to suppress NO x generate;

[0021] At the same time, combined with the O2 concentration data, when the O2 concentration is higher than the preset oxygen concentration threshold, the air supply volume is reduced to avoid excessive air causing NO x Increased generation;

[0022] When the CO concentration increases, it indicates incomplete combustion. The edge computing unit adjusts the air-coal ratio through the fuzzy PID algorithm, increasing the air volume or reducing the coal volume, while increasing the furnace temperature to promote complete combustion of CO.

[0023] When the dust concentration exceeds the standard, start the dust removal equipment to strengthen the operation and check the combustion status. However, if the dust increases due to unstable combustion, adjust the combustion parameters to make the combustion stable.

[0024] As a preferred solution of the multi-source sensing-based intelligent optimization method for coal-fired boilers described in the present invention, the dynamic adjustment based on the implementation results of the control strategy includes fuzzy PID control adjustment of the air-coal ratio and coordinated optimization adjustment of the burnout air swing angle and oxygen content;

[0025] The air-coal ratio fuzzy PID control is adjusted to use the O2 content and CO concentration detected in the tail flue as input variables;

[0026] The fuzzy PID controller adjusts the PID parameters including proportional coefficient, integral time, and differential time in real time through fuzzy reasoning according to the deviation and deviation change rate of the input variables, thereby correcting the air-coal ratio.

[0027] When the O2 concentration is high and the CO concentration is low, it means that the air volume is large, so reduce the air volume and the air-coal ratio;

[0028] When the O2 concentration is low and the CO concentration is high, it means that the air volume is small. Increase the air volume and increase the air-coal ratio.

[0029] The coordinated optimization of the burnout air swing angle and the oxygen content is to establish a correlation model between the burnout air swing angle and the oxygen content, and use the burnout air swing angle and the O2 concentration as optimization variables;

[0030] Using reinforcement learning algorithm to maximize combustion efficiency and NO x With the goal of minimizing emissions, we continuously learn and optimize during operation to obtain the optimal combination of burnout air swing angle and O2 concentration under different operating conditions.

[0031] The calculation unit sends instructions to the burnout air swing angle driver and the blower according to the optimization result to coordinate the burnout air swing angle driver and the blower, and the adjustment response time does not exceed 5 seconds.

[0032] The beneficial effects of the preferred technical solution in the embodiment of the present invention are: according to NO x The PID parameters are adjusted in real time based on the CO deviation change rate to solve the instability problem of conventional PID under variable load conditions; it breaks through the limitations of empirical formulas and independently seeks optimization under different working conditions to continuously improve boiler efficiency.

[0033] As a preferred solution of the multi-source sensing-based intelligent optimization method for coal-fired boilers described in the present invention, the coking warning includes predicting the contamination trend of the heating surface based on the LSTM neural network and providing an early warning of the coking risk;

[0034] The LSTM neural network model inputs historical and real-time data on furnace temperature field distribution, flue gas flow rate, and coal ash melting point;

[0035] The model predicts the contamination trend of the heated surface every 30 minutes. If a coking risk is predicted to occur in the next two hours, a red warning heat map of the coking risk will be displayed on the visual human-machine interface, and an audible and visual alarm will be issued.

[0036] Automatically recommend sootblowing timing. The edge computing unit sends instructions to the sootblowing equipment control module to start the sootblowing operation according to the recommended time.

[0037] Adjust the combustion temperature field. Based on the three-dimensional temperature field model, adjust the secondary air door opening or the burnout air swing angle in the area where coking may occur to reduce the temperature in the current area and avoid coking formation.

[0038] The beneficial effects of the preferred technical solution in the embodiment of the present invention are: the false alarm rate is reduced compared with the single temperature warning model, leaving operation time for active intervention, and traditional soot blowing only responds afterwards; the high-risk areas are accurately cooled to avoid efficiency loss caused by global cooling; the soot blowing cycle is dynamically recommended in combination with coking prediction, which reduces ineffective soot blowing and extends the life of the heated surface.

[0039] Another object of the present invention is to provide a coal-fired boiler intelligent optimization system based on multi-source perception.

[0040] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent optimization system for coal-fired boilers based on multi-source perception, comprising: a multi-dimensional perception module, a compensation module, an optimization control module, an early warning module, and an intelligent control terminal;

[0041] The multi-dimensional perception module, the coal quality online analyzer, is arranged in the downstream of the coal feeder coal falling pipe;The furnace three-dimensional temperature field scanner adopts an infrared array and an ultrasonic positioning mode, and one group is arranged every 2 meters in the burner area;The tail flue multi-component detection unit detects a four-in-one sensor of NO x , CO, O2 and dust, and is arranged at the outlet of the air preheater.

[0042] The compensation module collects and feeds back the compensation of the coal data of the coal-fired boiler.

[0043] The optimization control module collects the furnace data to reconstruct the combustion field, formulates an optimization control strategy, constructs an optimization target, and performs multi-component collaborative control on the tail flue.

[0044] The early warning module dynamically adjusts based on the implementation result of the control strategy, and monitors in real time to perform coking early warning.

[0045] The intelligent control terminal comprises an edge computing unit, a built-in combustion optimization algorithm, a self-adaptive adjustment execution mechanism, a secondary air door opening degree fine adjustment motor and a burnout air swing angle driver;It also comprises a visual man-machine interface, a three-dimensional thermal field dynamic display and a coking risk thermal map early warning.

[0046] The application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the intelligent optimization method for a coal-fired boiler based on multi-source perception when executing the computer program.

[0047] The application provides a computer readable storage medium, which stores a computer program, and wherein the computer program implements the steps of the intelligent optimization method for a coal-fired boiler based on multi-source perception when being executed by a processor.

[0048] The application has the following beneficial effects: through real-time coal quality detection, gamma-ray online analysis, 30 seconds / second update, dynamic air-coal ratio compensation, high-precision three-dimensional combustion field reconstruction, multi-component collaborative control of the tail flue, 2-hour early warning of LSTM coking prediction, a full-closed-loop optimization system, combustion efficiency is improved, coal quality-air volume dynamic matching reduces incomplete combustion loss, and boiler thermal efficiency is improved;Ultra-low emission is stable, NOx is stable and ammonia escape is reduced through burnout air swing angle reinforcement learning optimization;Intelligent prevention and control are upgraded, the three-dimensional temperature field regulation and control with a spatial resolution of 2 meters and LSTM coking prediction reduce invalid blowing and reduce power supply coal consumption. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 A flowchart of a coal-fired boiler intelligent optimization method based on multi-source perception is provided in accordance with an embodiment of the present invention.

[0051] Figure 2 An embodiment of the present invention provides an overall structure of a coal-fired boiler intelligent optimization method based on multi-source perception.

[0052] Figure 3 An embodiment of the present invention provides a coking warning method process for a coal-fired boiler intelligent optimization method based on multi-source perception. DETAILED DESCRIPTION

[0053] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0054] Example 1, with reference to Figures 1 to 3 , is an embodiment of the present invention, which provides a coal-fired boiler intelligent optimization method based on multi-source perception, including:

[0055] S100: Collect coal data for coal-fired boilers and perform feedback compensation.

[0056] S200: Collect furnace data to reconstruct the combustion field, formulate optimization control strategies, and construct optimization targets for multi-component coordinated control of the tail flue.

[0057] S300: Dynamically adjust based on the implementation results of the control strategy, and conduct real-time monitoring to provide early warning of coking.

[0058] It should be noted that the existing technology has the following problems: the traditional DCS system lacks multi-parameter coupling analysis, the existing optimization system does not consider the impact of coal quality fluctuations on combustion, and there is a lack of effective early warning measures for coking problems caused by local overheating.

[0059] Therefore, in response to the above-mentioned problems, a coal-fired boiler intelligent optimization system based on multi-source perception and its collaborative control method are designed through steps S100-S300; through the deep integration of multi-dimensional real-time perception (coal quality, furnace temperature field, tail flue gas components), edge intelligent computing and advanced control strategies, the three-in-one collaborative optimization of coal-fired boiler combustion efficiency, environmental emissions and operation safety is achieved.

[0060] Example 2, reference Figures 1 to 3 , is an embodiment of the present invention, which provides a coal-fired boiler intelligent optimization method based on multi-source perception, including:

[0061] In an embodiment of the present invention, the coal data of the coal-fired boiler is collected and feedback compensation is performed in S100, including the following steps S101-S102:

[0062] In an embodiment of the present invention, S101, installing a collection device, collecting furnace coal data and performing a detection input edge calculation to obtain coal type characteristic parameters, setting an update time to update the coal type characteristic parameters, includes the following steps A1-A2:

[0063] A1. Install the coal quality online analyzer in the middle and downstream of the coal feeder coal drop pipe;

[0064] A2. Using gamma ray technology, key parameters such as calorific value, ash content, moisture, and volatile matter of the coal entering the furnace are tested every 30 seconds, and the data is transmitted to the calculation unit in real time.

[0065] In an optional embodiment, the updating of coal characteristic parameters in S101 may be performed by installing a near-infrared spectroscopy probe above the coal feeder belt conveyor section, directly aiming it at the surface of the flowing coal, scanning the coal flow surface using near-infrared spectroscopy, collecting spectral data every 45 seconds, and calculating the calorific value, moisture content, volatile matter and other parameters of the coal through inversion using a built-in model.

[0066] In another optional embodiment, the updating of coal type characteristic parameters in S101 may also be performed by installing microwave moisture sensors on both sides of the coal drop pipe and installing a coal flow density detection device at the bottom of the pipe. The microwave sensor detects the moisture content of the coal in real time (at intervals of 40 seconds), and the density sensor estimates the ash ratio through pressure feedback, and calculates the calorific value and volatile matter in combination with the preset coal type empirical formula.

[0067] In an embodiment of the present invention, S102, dynamically correcting the air-coal ratio reference value for feedback compensation based on the coal characteristic parameters acquired in real time, includes the following steps B1-B4:

[0068] B1. The calculation unit dynamically corrects the air-coal ratio reference value based on the coal characteristic parameters obtained in real time.

[0069] B2. When the volatile matter of coal is high, in order to avoid deflagration, the primary air volume is automatically reduced. The reduction ratio is determined according to the extent to which the volatile matter exceeds the standard value. For every 1% increase in the extent, the primary air volume is reduced by 0.5%-1%.

[0070] B3. When the ash content of coal is high, in order to reduce incomplete combustion, the excess air coefficient is automatically increased. The increase ratio depends on the degree to which the ash content exceeds the standard value. For every 1% increase in ash content, the excess air coefficient increases by 0.2%-0.5%.

[0071] B4. The calculation unit sends the corrected air-coal ratio reference value to the control modules of the coal feeder and the blower to achieve real-time adjustment of the air-coal volume, with the adjustment response time controlled within 10 seconds.

[0072] In an optional embodiment, feedback compensation in S102 can be performed by manually calculating the correction to the baseline air-to-coal ratio based on the average coal quality obtained over a period of time and referring to an empirical table. If the volatile matter content exceeds the design value by X%, the baseline primary air volume value is adjusted downward by a fixed percentage of 5%; if the ash content exceeds the design value by Y%, the baseline excess air coefficient value is adjusted upward by a fixed percentage of 3%. The operator manually enters the corrected baseline air-to-coal ratio value on the DCS or control interface. The DCS system sends the manually entered correction value to the control modules of the coal feeder and blower for adjustment. The adjustment response time depends on the speed of manual operation and is typically much longer than 10 seconds. However, it cannot cope with rapid fluctuations in coal quality, has low compensation accuracy, and relies heavily on manual operation efficiency.

[0073] In another optional embodiment, the feedback compensation in S102 can also be that the calculation unit directly calls the corresponding fixed compensation value from the pre-stored empirical compensation value list according to the preset coal type (rather than the real-time parameter); if it is identified as high-volatile coal, the primary air volume baseline value is fixedly reduced by 4%; if it is identified as high-ash coal, the excess air coefficient baseline value is fixedly increased by 2.5%; if it is identified as the design coal type, the compensation value is zero, and the calculation unit applies the fixed compensation value to correct the air-coal ratio baseline value. The calculation unit sends the corrected air-coal ratio baseline value to the control module of the coal feeder and the blower for adjustment. It cannot perform fine adjustment according to the continuous changes in coal quality parameters, and the coal type identification is fuzzy and the compensation value is fixed.

[0074] In summary, the γ-ray coal quality online analyzer can detect the key parameters of the coal entering the furnace, such as calorific value, ash content, moisture, volatile matter, etc. in real time (30 seconds / time), and dynamically correct the reference value of the air-coal ratio, so as to achieve the goal of reducing the air flow to prevent explosion of high-volatile coal and increasing the air flow to promote combustion of high-ash coal. The closed-loop control is completed within 10 seconds, solving the problems of lag in traditional manual sampling and inaccurate air-coal ratio, thus improving combustion efficiency and NO x Volatility is reduced.

[0075] In an embodiment of the present invention, in S200, furnace data is collected to reconstruct the combustion field, an optimization control strategy is formulated, and an optimization target is constructed to perform multi-component coordinated control of the tail flue, including the following steps S201-S204:

[0076] In an embodiment of the present invention, S201, collecting furnace data to reconstruct the combustion field, includes the following steps C1-C2:

[0077] C1. The furnace three-dimensional temperature field scanner adopts infrared array and ultrasonic positioning method. A group is arranged every 2 meters in the burner area to collect temperature information and spatial coordinate data at different positions in the furnace in real time and transmit it to the calculation unit.

[0078] C2. Using the collected temperature and coordinate data, we reconstruct the furnace's three-dimensional temperature field model every 5 minutes, ensuring that the error between the model and the actual combustion state is less than 5°C. This model accurately identifies high-temperature and low-temperature areas, as well as localized biased burning.

[0079] In an optional embodiment, the combustion field reconstruction in S201 may be:

[0080] In the furnace burner area, a grid is divided along the height and width directions. A group of armored thermocouples is fixedly installed between the water-cooled wall tubes of the furnace wall in each unit area. Each group contains at least three thermocouples with different insertion depths. The computing unit collects the temperature readings of all thermocouples in real time. Using the position coordinates and measured temperature values ​​of the thermocouples, combined with the empirical furnace heat transfer model, a spatial interpolation algorithm (such as the inverse distance weighted method or Kriging interpolation) is used to reconstruct the three-dimensional temperature field of the entire furnace.

[0081] In another optional embodiment, the combustion field reconstruction in S201 can also be performed by fixedly installing multiple groups of sound wave transmitters and receivers at relative positions on the four walls of the furnace to form a sound wave propagation path network that cross-covers the cross-section of the furnace; the computing unit controls the transmitters to sequentially transmit sound wave pulses of a specific frequency, and synchronously records the time when each receiver receives the signal. The propagation speed of the sound wave in the flue gas is directly related to the flue gas temperature. The higher the temperature, the greater the sound speed; the computing unit collects the propagation time data of all sound wave paths, and uses the tomographic imaging inversion algorithm to calculate the two-dimensional / three-dimensional temperature distribution of the area covered by the path network.

[0082] S202. Based on the results of the three-dimensional temperature field model analysis, if there is a local high-temperature area, the edge computing unit sends a command to the secondary air damper opening fine-tuning motor to appropriately reduce the secondary air damper opening corresponding to the area, thereby reducing the oxygen supply and thus reducing the local temperature. The opening adjustment accuracy reaches ±1%;

[0083] If there is a low-temperature area or localized uneven burning, increase the secondary air door opening in the corresponding area to increase the oxygen supply and disturbance intensity to promote uniform combustion.

[0084] S203, a multi-component detection unit is installed at the outlet of the air preheater in the tail flue, and a four-in-one sensor is used to monitor NO in real time. x , CO, O2, and dust concentrations, and feeds the data back to the edge computing unit in real time;

[0085] When NO x Emission concentration is less than the emission threshold of 50mg / Nm 3 When the burnout wind angle adjustment mechanism is activated, the edge computing unit adjusts the burnout wind angle within the range of ±30 degrees through the burnout wind angle driver to suppress NO x generate;

[0086] At the same time, combined with the O2 concentration data, when the O2 concentration is higher than the preset oxygen concentration threshold, the air supply volume is reduced to avoid excessive air causing NO x Increased generation;

[0087] When the CO concentration increases, it indicates incomplete combustion. The edge computing unit adjusts the air-coal ratio through the fuzzy PID algorithm, increasing the air volume or reducing the coal volume, while increasing the furnace temperature to promote complete combustion of CO.

[0088] When the dust concentration exceeds the standard, start the dust removal equipment to strengthen the operation and check the combustion status. However, if the dust increases due to unstable combustion, adjust the combustion parameters to make the combustion stable.

[0089] S204, dynamic adjustment based on the implementation results of the control strategy includes fuzzy PID control adjustment of the air-coal ratio and coordinated optimization adjustment of the burnout air swing angle and oxygen content;

[0090] The air-coal ratio fuzzy PID control is adjusted to use the O2 content and CO concentration detected in the tail flue as input variables;

[0091] The fuzzy PID controller adjusts the PID parameters including proportional coefficient, integral time, and differential time in real time through fuzzy reasoning according to the deviation and deviation change rate of the input variables, thereby correcting the air-coal ratio.

[0092] When the O2 concentration is high and the CO concentration is low, it means that the air volume is large, so reduce the air volume and the air-coal ratio;

[0093] When the O2 concentration is low and the CO concentration is high, it means that the air volume is small. Increase the air volume and increase the air-coal ratio.

[0094] The coordinated optimization of the burnout air swing angle and the oxygen content is to establish a correlation model between the burnout air swing angle and the oxygen content, and use the burnout air swing angle and the O2 concentration as optimization variables;

[0095] Using reinforcement learning algorithm to maximize combustion efficiency and NO xWith the goal of minimizing emissions, we continuously learn and optimize during operation to obtain the optimal combination of burnout air swing angle and O2 concentration under different operating conditions.

[0096] The calculation unit sends instructions to the burnout air swing angle driver and the blower according to the optimization result to coordinate the burnout air swing angle driver and the blower, and the adjustment response time does not exceed 5 seconds.

[0097] In summary, based on infrared + ultrasonic three-dimensional temperature field scanning, the secondary air door opening is dynamically adjusted with an accuracy of ±1% to balance the furnace combustion; combined with the tail flue NO x , CO, O2, and dust real-time data, using fuzzy PID to control the air-coal ratio, and through reinforcement learning to optimize the burnout air swing angle and oxygen content, responding to changes within 5 seconds to achieve NO x Stable, reduced fly ash carbon content, while reducing dust removal energy consumption.

[0098] In an embodiment of the present invention, dynamic adjustment is performed based on the implementation results of the control strategy in S300, and real-time monitoring is performed to provide a coking warning, including the following steps S301-S302:

[0099] In an embodiment of the present invention, S301, real-time monitoring and early warning of coking, includes the following steps D1-D2:

[0100] D1. The LSTM neural network model inputs historical and real-time data on furnace temperature field distribution, flue gas flow rate, and coal ash melting point.

[0101] D2. The model predicts the contamination trend of the heated surface every 30 minutes. If a coking risk is predicted to occur in the next two hours, a red warning heat map of the coking risk will be displayed on the visual human-machine interface, and an audible and visual alarm will be issued.

[0102] In an optional embodiment, the coking warning in S301 can be to install a high-density thermocouple array (one measuring point per 1 meter × 1 meter grid) in the key areas of the furnace water-cooled wall to collect wall temperature data in real time; synchronously access the real-time ash melting point data provided by the coal quality analyzer, and set the basic warning temperature threshold = real-time ash melting point - 50°C; when the temperatures of three consecutive measuring points in a certain area are greater than the basic warning temperature threshold, and the temperature difference between adjacent measuring points is greater than 100°C / m, it is determined to be local overheating; if the local overheating state lasts for 10 minutes, a yellow warning area is marked on the human-machine interface; if the local temperature rises to -20°C of the ash melting point and lasts for 5 minutes, a red sound and light alarm is triggered.

[0103] In another optional embodiment, the coking warning in S301 can also be to install a laser scattering ash particle concentration meter at the smoke window of the furnace outlet to monitor the ash content of the flue gas in real time, and set a weighing ash accumulation detection plate at the air preheater inlet (automatically weighed and reset every 4 hours). According to the real-time coal ash content, flue gas flow rate and ash content, the theoretical ash deposition rate is calculated, and the actual weight increase of the ash accumulation detection plate is compared to calculate the deposition deviation coefficient K (K = measured value / theoretical value). When K>1.5 and lasts for 2 hours, it is determined to be abnormal ash accumulation acceleration. If the wall temperature of the area is monitored to be>500℃ at the same time, the coking warning signal is triggered.

[0104] S302: Automatically recommend a sootblowing time. The edge computing unit sends an instruction to the sootblowing equipment control module to start the sootblowing operation according to the recommended time.

[0105] Adjust the combustion temperature field. Based on the three-dimensional temperature field model, adjust the secondary air door opening or the burnout air swing angle in the area where coking may occur to reduce the temperature in the current area and avoid coking formation.

[0106] In summary, the LSTM model is used to fuse temperature field, smoke velocity, and ash melting point data to predict the coking risk in the next two hours every 30 minutes. The local temperature is adjusted through the secondary air damper / burnout air, and the timing of soot blowing is intelligently recommended to reduce ineffective soot blowing, lower the false alarm rate of coking, avoid damage to the heating surface, and extend the boiler operation cycle.

[0107] Based on the solution of Example 2, the present invention can achieve a 0.8-1.2% improvement in boiler efficiency and a 15-20 mg / Nm reduction in NOx emissions for a 600MW unit. 3 , reducing the frequency of manual adjustments by more than 80%, the accuracy of coking warning reaches 92%, and saving about 2 million yuan in coal costs annually.

[0108] Example 3 is an embodiment of the present invention. The above is a schematic scheme of a method for intelligent optimization of a coal-fired boiler based on multi-source sensing. It should be noted that the technical solution of a system for intelligent optimization of a coal-fired boiler based on multi-source sensing and the technical solution of the method for intelligent optimization of a coal-fired boiler based on multi-source sensing are based on the same concept. For details not described in detail in the technical solution of the system for intelligent optimization of a coal-fired boiler based on multi-source sensing in this embodiment, please refer to the description of the technical solution of the method for intelligent optimization of a coal-fired boiler based on multi-source sensing.

[0109] This embodiment also provides a coal-fired boiler intelligent optimization system based on multi-source perception, including: a multi-dimensional perception module, a compensation module, an optimization control module, an early warning module, and an intelligent control terminal;

[0110] The multi-dimensional sensing module and coal quality online analyzer are set in the middle and downstream of the coal feeder coal drop pipe; the furnace three-dimensional temperature field scanner adopts infrared array and ultrasonic positioning mode, and a group is arranged every 2 meters in the burner area; the tail flue multi-component detection unit detects NO x , CO, O2 dust four-in-one sensor, air preheater outlet arrangement.

[0111] The compensation module collects coal data from coal-fired boilers and performs feedback compensation;

[0112] The optimization control module collects furnace data to reconstruct the combustion field, formulates optimization control strategies, and constructs optimization targets to perform multi-component coordinated control of the tail flue;

[0113] The early warning module is dynamically adjusted based on the implementation results of the control strategy, and performs real-time monitoring to provide early warning of coking;

[0114] The intelligent control terminal includes an edge computing unit, a built-in combustion optimization algorithm, an adaptive adjustment actuator, a secondary air door opening fine-tuning motor, and a burnout air swing angle driver; it also includes a visual human-machine interface, a three-dimensional thermal field dynamic display, and a coking risk thermal map warning.

[0115] This embodiment also provides an electronic device, which is suitable for a coal-fired boiler intelligent optimization method based on multi-source perception, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement a coal-fired boiler intelligent optimization method based on multi-source perception as proposed in the above embodiment.

[0116] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, an intelligent optimization method for a coal-fired boiler based on multi-source perception as proposed in the above embodiment is implemented.

[0117] The storage medium proposed in this embodiment and the method for realizing an intelligent optimization method of a coal-fired boiler based on multi-source perception proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0118] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and necessary universal hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a ROM, a RAM, a FLASH, a hard disk, or an optical disc, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present application.

[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A coal-fired boiler intelligent optimization method based on multi-source perception, characterized by: include, Collect coal data for coal-fired boilers and provide feedback compensation; Collect furnace data to reconstruct the combustion field, formulate optimization control strategies, and establish optimization targets for multi-component coordinated control of the tail flue; Dynamic adjustments are made based on the implementation results of the control strategy, and real-time monitoring is carried out to provide early warning of coking.

2. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 1, characterized in that: The feedback compensation includes: Install a collection device to collect furnace coal data and perform a detection input into edge computing to obtain coal type characteristic parameters; Set the update time to update the coal characteristic parameters; According to the coal characteristic parameters obtained in real time, the air-coal ratio benchmark value is dynamically corrected to compensate for coal quality.

3. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 2, characterized in that: The combustion field reconstruction comprises: Use positioning method to re-collect furnace data and input it into edge computing; Set the reconstruction time based on the furnace data and reconstruct the three-dimensional temperature field model of the furnace; Different temperature zones are identified through a three-dimensional temperature field model.

4. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 3, characterized in that: The formulation of the optimization control strategy includes: Based on the results of the three-dimensional temperature field model identification and analysis, if there is a high-temperature area, the edge computing sends a cooling instruction to adjust the opening accuracy; If there is a low temperature area or localized burning, increase the opening.

5. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 4, characterized in that: The optimization target is constructed to carry out the coordinated control of the tail flue multi-components, including installing a multi-component detection unit at the outlet of the air preheater, and monitoring the NO in real time through a four-in-one sensor. x , CO, O2, and dust concentrations, and feeds the data back to the edge computing unit in real time; When NO x Emission concentration is less than the emission threshold of 50mg / Nm 3 When the burnout wind angle adjustment mechanism is activated, the edge computing unit adjusts the burnout wind angle within the range of ±30 degrees through the burnout wind angle driver to suppress NO x generate; At the same time, combined with the O2 concentration data, when the O2 concentration is higher than the preset oxygen concentration threshold, the air supply volume is reduced to avoid excessive air causing NO x Increased generation; When the CO concentration increases, it indicates incomplete combustion. The edge computing unit adjusts the air-coal ratio through the fuzzy PID algorithm, increasing the air volume or reducing the coal volume, while increasing the furnace temperature to promote complete combustion of CO. When the dust concentration exceeds the standard, start the dust removal equipment to strengthen the operation and check the combustion status. However, if the dust increases due to unstable combustion, adjust the combustion parameters to make the combustion stable.

6. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 5, characterized in that: The dynamic adjustment based on the implementation results of the control strategy includes fuzzy PID control adjustment of the air-coal ratio and coordinated optimization adjustment of the burnout air swing angle and oxygen content; The air-coal ratio fuzzy PID control is adjusted to use the O2 content and CO concentration detected in the tail flue as input variables; The fuzzy PID controller adjusts the PID parameters including proportional coefficient, integral time, and differential time in real time through fuzzy reasoning according to the deviation and deviation change rate of the input variables, thereby correcting the air-coal ratio. When the O2 concentration is high and the CO concentration is low, it means that the air volume is large, so reduce the air volume and the air-coal ratio; When the O2 concentration is low and the CO concentration is high, it means that the air volume is small. Increase the air volume and increase the air-coal ratio. The coordinated optimization of the burnout air swing angle and the oxygen content is to establish a correlation model between the burnout air swing angle and the oxygen content, and use the burnout air swing angle and the O2 concentration as optimization variables; Using reinforcement learning algorithm to maximize combustion efficiency and NO x With the goal of minimizing emissions, we continuously learn and optimize during operation to obtain the optimal combination of burnout air swing angle and O2 concentration under different operating conditions. The calculation unit sends instructions to the burnout air swing angle driver and the blower according to the optimization result to coordinate the burnout air swing angle driver and the blower, and the adjustment response time does not exceed 5 seconds.

7. The method for intelligent optimization of a coal-fired boiler based on multi-source sensing according to claim 6, characterized in that: The coking warning includes predicting the contamination trend of the heated surface based on the LSTM neural network and providing early warning of coking risks; The LSTM neural network model inputs historical and real-time data on furnace temperature field distribution, flue gas flow rate, and coal ash melting point; The model predicts the contamination trend of the heated surface every 30 minutes. If a coking risk is predicted to occur in the next two hours, a red warning heat map of the coking risk will be displayed on the visual human-machine interface, and an audible and visual alarm will be issued. Automatically recommend sootblowing timing. The edge computing unit sends instructions to the sootblowing equipment control module to start the sootblowing operation according to the recommended time. Adjust the combustion temperature field. Based on the three-dimensional temperature field model, adjust the secondary air door opening or the burnout air swing angle in the area where coking may occur to reduce the temperature in the current area and avoid coking formation.

8. A coal-fired boiler intelligent optimization system based on multi-source perception, applying the coal-fired boiler intelligent optimization method based on multi-source perception according to any one of claims 1 to 7, characterized in that: include: Multi-dimensional perception module, compensation module, optimization control module, early warning module, and intelligent control terminal; The multi-dimensional sensing module and coal quality online analyzer are set in the middle and downstream of the coal feeder coal drop pipe; the furnace three-dimensional temperature field scanner adopts infrared array and ultrasonic positioning mode, and a group is arranged every 2 meters in the burner area; the tail flue multi-component detection unit detects NO x , CO, O2 dust four-in-one sensor, air preheater outlet arrangement. The compensation module collects coal data from coal-fired boilers and performs feedback compensation; The optimization control module collects furnace data to reconstruct the combustion field, formulates optimization control strategies, and constructs optimization targets to perform multi-component coordinated control of the tail flue; The early warning module is dynamically adjusted based on the implementation results of the control strategy, and performs real-time monitoring to provide early warning of coking; The intelligent control terminal includes an edge computing unit, a built-in combustion optimization algorithm, an adaptive adjustment actuator, a secondary air door opening fine-tuning motor, and a burnout air swing angle driver; it also includes a visual human-machine interface, a three-dimensional thermal field dynamic display, and a coking risk thermal map warning.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the processor implements the steps of a coal-fired boiler intelligent optimization method based on multi-source perception according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a coal-fired boiler intelligent optimization method based on multi-source perception according to any one of claims 1 to 7 are implemented.

Images