Boiler heating surface slagging monitoring system and method based on convolutional neural network algorithm
By setting up laser absorption spectroscopy measurement modules and convolutional neural network algorithms in different sections of the boiler, the problem of monitoring slagging inside the boiler furnace was solved, enabling real-time monitoring and soot blowing optimization of the boiler's heating surfaces, thus improving the safety and economy of boiler operation.
Patent Information
- Application Number
- CN202410952536.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
Existing technologies cannot achieve online monitoring and accurate diagnosis of slagging conditions inside the boiler furnace, resulting in a lack of precision in soot blowing operations and easily causing damage to the water-cooled walls or severe slagging.
A boiler heating surface slagging monitoring system based on convolutional neural network algorithm is adopted. By setting up laser absorption spectroscopy measurement modules in the main combustion zone, burnout zone and heat exchange zone, and combining adaptive iterative algorithm and convolutional neural network, a two-dimensional temperature field and CO gas concentration field are reconstructed, and an ash slagging monitoring model is established to realize real-time monitoring and soot blowing optimization of different heating surfaces.
It enables intuitive acquisition of the ash and slag accumulation status of different heating surfaces inside the boiler, optimizes the soot blowing frequency, and improves the safety and economy of boiler operation.
Smart Images

Figure CN118914131A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of flame combustion diagnosis, and in particular to a boiler heating surface slagging monitoring system and method based on a convolutional neural network algorithm. Background Art
[0002] Thermal power plants use coal as fuel. Since the ash and sulfur content of coal are relatively high, the combustion products will form stains on the heating surface. In the actual operation of the boiler, the slagging problem of the boiler is inevitable. The hazards of slagging are mainly: serious ash and slagging in the furnace will increase the heat transfer resistance, resulting in a decrease in the heat absorption of the water-cooled wall, an increase in the exhaust gas temperature, and a decrease in the boiler efficiency; ash and slagging on the heating surface will cause abrasion, corrosion and thinning of the furnace tube surface; with the increasing demand for low-load operation of the unit, the reduction of the flue gas temperature and flow rate in the furnace will cause the slag blocks in the furnace to fall off and the ash blockage of the convection heating surface will increase. These factors reduce the safety of the boiler under low-load operation, increase coal consumption, and seriously damage the economic benefits. When the slagging phenomenon is serious, it may even cause the unit to operate at a reduced load, and the frequent falling of large slag blocks may even cause accidents such as equipment damage. Therefore, the safety of the heating surface of the pulverized coal boiler and the protection of ash and slag need to be given extreme attention.
[0003] At present, the commonly used and effective ash removal and slag removal technology is to perform soot blowing during the operation of the boiler. Soot blowing is mostly performed at a fixed time and fixed duration according to manual experience, without considering the actual contamination of the boiler heating surface. It is highly subjective and prone to improper soot blowing. Too frequent soot blowing will cause damage to the water-cooled wall and loss of unit efficiency; while insufficient soot blowing will cause serious slagging of the water-cooled wall. Therefore, monitoring the ash accumulation and slagging of the boiler and achieving precise soot blowing have become the main trends in soot blowing in the furnace of power plants. The online monitoring of the ash accumulation on the convection heating surface of the boiler is relatively mature, mainly using the balance method. Generally, the required measurement data can be measured through conventional measurement points in the power plant. The boiler slagging mainly occurs on the radiation heating surface inside the furnace. Due to the harsh environment inside the furnace, the combustion of pulverized coal involves a variety of physical and chemical reactions, the situation is very complicated, and there is a lack of suitable measuring devices, which makes the monitoring and diagnosis of furnace slagging difficult. In order to judge the degree and location of slagging inside the furnace as accurately as possible, domestic scholars have conducted research and analysis on the monitoring and diagnosis methods of furnace slagging. At present, the main slagging monitoring methods include acoustic temperature measurement, infrared radiation temperature measurement, monitoring of furnace outlet smoke temperature, heat flux meter method, furnace thermal effectiveness coefficient method, water-cooled wall fin end temperature characterization method and other methods.
[0004] Most of the above methods can only diagnose the overall slagging condition of the furnace. In fact, the temperature and heat load in the furnace are uneven, and the contamination of the heating surface is also different. Therefore, the existing technology cannot achieve online monitoring and accurate diagnosis of the slagging condition of the furnace. Summary of the invention
[0005] In order to solve the technical problem that the coking and slagging phenomenon on the heating surface of the pulverized coal boiler in the power station cannot be accurately diagnosed in the prior art, the present invention provides a boiler heating surface slagging monitoring system and method based on the convolutional neural network algorithm, which directly monitors the ash accumulation dynamics by optical means, and combines effective optical path arrangement and intelligent algorithm to intuitively obtain the ash accumulation and slagging status of each heating surface in the main combustion zone, burnout zone and heat exchange zone in the furnace, realize real-time monitoring of the temperatures of different heating surfaces in the furnace, and optimize the over-temperature warning of the furnace tube and the guidance of the soot blowing frequency.
[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a boiler heating surface slagging monitoring system based on a convolutional neural network algorithm, comprising: a computing unit and three laser absorption spectrum measurement modules respectively arranged in a main combustion zone, a burnout zone and a heat exchange zone, the laser absorption spectrum measurement module comprising M first measurement optical paths, N second measurement optical paths and corresponding detectors, wherein the M first measurement optical paths and the N second measurement optical paths are arranged vertically and located in the same horizontal plane, and the M first measurement optical paths and the N second measurement optical paths divide the measurement area into an M×N grid; the detector is used to measure the transmission spectrum of each measurement optical path;
[0007] The calculation unit includes a two-dimensional field reconstruction unit, a deep learning unit and a soot blowing control unit. The output end of the detector is connected to the two-dimensional field reconstruction unit through an acquisition card. The two-dimensional field reconstruction unit is used to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively according to the spectral information obtained by the detector in each laser absorption spectrum measurement module;
[0008] The deep learning unit is provided with a boiler water-cooled wall ash and slagging monitoring model established by a convolutional neural network, and the boiler water-cooled wall ash and slagging monitoring model is used to calculate the ash deposition parameters of the main combustion zone, the burnout zone and the heat exchange zone according to the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively;
[0009] The soot blowing control unit is used to control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively according to the soot accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone.
[0010] The input parameters of the boiler water-cooled wall ash and slagging monitoring model include two-dimensional temperature field and CO concentration distribution, boiler load, coal quality parameters, and initial flue gas temperature at the furnace outlet, and the output parameters are dynamic characteristic parameters.
[0011] The boiler heating surface slagging monitoring system based on the convolutional neural network algorithm also includes: a furnace outlet flue gas temperature measuring device, which is used to obtain the flue gas temperature at the furnace outlet and send it to the calculation unit; the calculation unit is connected to the electric field DCS system, and is used to obtain coal quality parameters and boiler load from the electric field DCS system.
[0012] The boiler heating surface slagging monitoring system based on the convolutional neural network algorithm also includes soot blowers respectively arranged in the main combustion zone, the burnout zone and the heat exchange zone, and the control end of the soot blower is connected to the soot blowing control unit.
[0013] The dust accumulation parameter is specifically a dynamic characteristic parameter, and the definition formula of the dynamic characteristic parameter is:
[0014]
[0015] Among them, C represents the dynamic characteristic parameters of each region, T a Represents the average furnace temperature in each area, T b ' i Represents the surface temperature distribution value of the slag layer near the wall in the two-dimensional discrete grid of each area.
[0016] The specific method of the sootblowing control unit controlling the main combustion zone, the burnout zone and the heat exchange zone to perform sootblowing operations respectively according to the dynamic characteristic parameters of the main combustion zone, the burnout zone and the heat exchange zone is:
[0017] It is determined whether the dynamic characteristic parameters of each zone are less than a threshold value. If so, it is determined that slagging occurs in the zone, and the soot blower in the corresponding zone is controlled to perform soot blowing operation.
[0018] The laser absorption spectrum measurement module also includes a laser controller, a signal generator, a laser, a fiber optic beam splitter and a plurality of collimators. The signal generator is used to input a sawtooth wave signal to the laser controller, and the laser controller is used to control the output laser frequency and scanning range of the laser according to the sawtooth wave signal. The laser output by the laser is divided into M+N beams of laser light after passing through the fiber optic beam splitter, and then passes through a collimator to form a corresponding first measurement optical path and a second measurement optical path.
[0019] The two-dimensional field reconstruction unit adopts an adaptive iterative algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively.
[0020] In addition, the present invention also provides a boiler heating surface slagging monitoring method based on a convolutional neural network algorithm, comprising the following steps:
[0021] Step 1: respectively set up laser absorption spectrum measurement modules in the main combustion zone, burnout zone and heat exchange zone;
[0022] Step 2: Obtain the transmission spectrum data of each area, and use the adaptive iterative algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion area, burnout area and heat exchange area respectively;
[0023] Step 3: Based on the convolutional neural network, a boiler water wall ash and slagging monitoring model is established, and the historical operation data of the power plant is obtained to train the model;
[0024] Step 4: Input the real-time two-dimensional temperature field and CO gas concentration field data of the main combustion zone, burnout zone and heat exchange zone into the trained model, and output the ash accumulation parameters of the main combustion zone, burnout zone and heat exchange zone respectively through the model;
[0025] Step 5: According to the ash accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone, control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively.
[0026] Compared with the prior art, the present invention has the following beneficial effects:
[0027] The present invention provides a boiler heating surface slagging monitoring system and method based on a convolutional neural network algorithm. By arranging laser absorption spectrum measurement modules in the main combustion zone, burnout zone and heat exchange zone of the boiler respectively, and then using the adaptive iteration (AART) algorithm, the temperature field and CO concentration field distribution of each part of the heating surface of different sections are first obtained; the artificial convolutional neural network algorithm is used to establish a power station boiler water-cooled wall ash slagging monitoring model. The constructed model takes the temperature and CO concentration distribution of different sections, boiler load, coal quality, and furnace outlet flue gas temperature as input parameters, and the dynamic characteristic parameters characterizing the pollution status of the furnace heating surface are used as network outputs. The ash slagging state of different sections in the furnace is measured by the dynamic characteristic parameters. Therefore, the present invention can intuitively obtain the ash slagging state of each heating surface in different sections of the furnace, realize real-time monitoring of the temperature of different heating surfaces in the furnace, and optimize the over-temperature warning of the furnace tube and the guidance of the soot blowing frequency. The present invention realizes the furnace segmented soot blowing according to the specific area of contamination in the furnace, and intelligently controls the soot blowing system, which is of great significance for improving the safety and economy of the power station boiler operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 A schematic diagram of the structure of a boiler heating surface slagging monitoring system based on a convolutional neural network algorithm provided in Embodiment 1 of the present invention;
[0029] Figure 2 This is a schematic diagram of the arrangement of the laser absorption spectrum measurement module on a boiler in the first embodiment of the present invention;
[0030] Figure 3 This is a schematic diagram of the structure of a laser absorption spectrum measurement module in an embodiment of the present invention;
[0031] Figure 4 A schematic flow chart of a method for monitoring slagging on a boiler heating surface based on a convolutional neural network algorithm provided in the second embodiment of the present invention;
[0032] In the figure: 1 is the first measuring optical path, 2 is the second measuring optical path, 3 is the detector, 4 is the main combustion zone, 5 is the burnout zone, 6 is the heat exchange zone, 7 is the cold ash hopper, 8 is the soot blower, 9 is the laser absorption spectrum measurement module, 10 is the measuring area, 11 is the laser, 12 is the optical fiber beam splitter, 13 is the collimator, and 14 is the calculation unit. DETAILED DESCRIPTION
[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0034] like Figure 1 As shown, the first embodiment of the present invention provides a boiler heating surface slagging monitoring system based on a convolutional neural network algorithm, including: a computing unit and three laser absorption spectrum measurement modules 9; Figure 2 As shown, the boiler from bottom to top includes a cold ash hopper 7, a main combustion zone 4, a burnout zone 5 and a heat exchange zone 6. Three laser absorption spectrum measurement modules 9 are respectively arranged in the main combustion zone 4, the burnout zone 5 and the heat exchange zone 6 to measure the transmission spectrum of each measurement area.
[0035] like Figure 3 As shown, in this embodiment, the laser absorption spectrum measurement module includes M first measurement light paths 1, N second measurement light paths 2 and corresponding detectors 3, wherein the M first measurement light paths 1 and the N second measurement light paths 2 are vertically arranged and located in the same horizontal plane, and the M first measurement light paths 1 and the N second measurement light paths 2 divide the measurement area 10 into an M×N grid; the detector 3 is used to measure the transmission spectrum of each measurement light path.
[0036] Specifically, in this embodiment, Figure 3As shown, the laser absorption spectrum measurement module also includes a laser controller, a signal generator, a laser 11, a fiber beam splitter 12 and a plurality of collimators 13. The signal generator is used to input a sawtooth wave signal to the laser controller. The laser controller is used to control the output laser frequency and scanning range of the laser 11 according to the sawtooth wave signal. The laser output by the laser 11 is divided into M+N beams after passing through the fiber beam splitter and then passes through a collimator to form a corresponding first measurement optical path and a second measurement optical path. The laser is specifically a CO laser. The laser output by the tunable laser diode is divided into several beams after being combined by the fiber beam splitter 12, and then sent to the collimators 13 arranged horizontally in two dimensions through several optical fibers. After passing through the collimator 13, the output forms the first measurement optical path 1 and the second measurement optical path 2.
[0037] Furthermore, in this embodiment, the laser absorption spectrum measurement module further includes a filter, which is arranged in front of the detector 3 and is used to filter out flame radiation to reduce the influence of flame radiation on the reconstruction measurement.
[0038] Specifically, Figure 1 As shown, the calculation unit includes a two-dimensional field reconstruction unit, a deep learning unit and a soot blowing control unit. The output end of the detector is connected to the two-dimensional field reconstruction unit through an acquisition card. The two-dimensional field reconstruction unit is used to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively according to the spectral information obtained by the detector in each laser absorption spectrum measurement module; the deep learning unit is provided with a boiler water-cooled wall ash accumulation and slagging monitoring model established by a convolutional neural network, and the boiler water-cooled wall ash accumulation and slagging monitoring model is used to calculate the dynamic characteristic parameters of the main combustion zone, the burnout zone and the heat exchange zone according to the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively; the soot blowing control unit is used to control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively according to the ash accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone.
[0039] Specifically, in this embodiment, the boiler water wall ash and slagging monitoring model is obtained by training from the historical operation data of the power plant.
[0040] Specifically, in this embodiment, the two-dimensional field reconstruction unit adopts an adaptive iteration (AART) algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, burnout zone and heat exchange zone, respectively. Since the boiler system is very complex and physical simulation is difficult to carry out, the clean heat absorption of the heating surface is affected by many factors and is a nonlinear problem. Among the many reconstruction algorithms, the AART algorithm is widely used in the reconstruction of combustion field temperature and concentration distribution because it can well reconstruct the parameter distribution in the two-dimensional field for incomplete projection data or when the projection data is not evenly distributed in the range of 180° or 360°. After reconstructing the two-dimensional temperature field and CO concentration field of each zone, the average temperature and average CO concentration value of the furnace cross section of the main combustion zone, burnout zone and heat exchange zone can be obtained by mean calculation.
[0041] Specifically, Figure 4 As shown, in this embodiment, the input parameters of the boiler water-cooled wall ash and slagging monitoring model include two-dimensional temperature field and CO concentration distribution, boiler load, coal quality parameters, and initial flue gas temperature at the furnace outlet, and the output parameter is the ash deposition parameter.
[0042] Specifically, in this embodiment, the ash deposition parameter output by the boiler water wall ash and slagging monitoring model is specifically a dynamic characteristic parameter, and the definition formula of the dynamic characteristic parameter is:
[0043]
[0044] Among them, C represents the dynamic characteristic parameters of each region, T a Represents the average furnace temperature in each area, T b ' i Represents the surface temperature distribution value of the slag layer near the wall in the two-dimensional discrete grid of each area.
[0045] In this embodiment, as the thickness of the slag on the heating surface increases, the temperature T b ' i The temperature of the furnace increases sharply and gradually approaches the average furnace temperature. The more serious the slagging is, The closer it is to 1, the closer the dynamic characteristic parameter C is to 0. Therefore, in this embodiment, the dynamic characteristic parameter C can well express the pollution status of each area in the boiler. When the dynamic characteristic parameter C is less than the threshold value, it means that the slagging in the area has reached a certain level, and the sootblowing control unit starts the sootblower in the area.
[0046] Furthermore, the present embodiment also includes: a furnace outlet flue gas temperature measuring device, wherein the furnace outlet flue gas temperature measuring device is used to obtain the flue gas temperature at the furnace outlet and send it to the calculation unit.
[0047] Furthermore, in this embodiment, the calculation unit is connected to the power plant DCS system to obtain coal quality parameters and boiler load from the power plant DCS system. Specifically, the coal quality parameters include: received basis lower calorific value, volatile matter, ash content, and moisture content of the coal.
[0048] Furthermore, if Figure 1 and 4 As shown, a boiler heating surface slagging monitoring system based on a convolutional neural network algorithm in this embodiment also includes soot blowers respectively arranged in the main combustion zone, the burnout zone and the heat exchange zone, and the control end of the soot blower is connected to the soot blowing control unit.
[0049] Specifically, the sootblowing control unit controls the main combustion zone, the burnout zone and the heat exchange zone to perform sootblowing operations respectively according to the dynamic characteristic parameters of the main combustion zone, the burnout zone and the heat exchange zone. The specific method is: determine whether the dynamic characteristic parameters of each zone are less than a threshold value; if so, determine that the area is slagging, and control the sootblower in the corresponding area to perform sootblowing operations.
[0050] Embodiment 2
[0051] like Figure 4 As shown, the second embodiment of the present invention provides a boiler heating surface slagging monitoring method based on a convolutional neural network algorithm, comprising the following steps:
[0052] Step 1: respectively set up laser absorption spectrum measurement modules in the main combustion zone, burnout zone and heat exchange zone;
[0053] Step 2: Obtain the transmission spectrum data of each area, and use the adaptive iterative algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion area, burnout area and heat exchange area respectively;
[0054] Step 3: Based on the convolutional neural network, a boiler water wall ash and slagging monitoring model is established, and the historical operation data of the power plant is obtained to train the model.
[0055] Specifically, the historical data includes the two-dimensional temperature field and CO concentration distribution obtained from historical measurements, boiler load, coal quality parameters, flue gas temperature at the furnace outlet, and corresponding ash accumulation parameters. The ash accumulation parameters are specifically dynamic characteristic parameters, which are obtained by calculating the historical data and formula (1).
[0056] Step 4: Input the real-time two-dimensional temperature field and CO gas concentration field data of the main combustion zone, burnout zone and heat exchange zone into the trained model, and output the ash accumulation parameters of the main combustion zone, burnout zone and heat exchange zone respectively through the model;
[0057] Step 5: According to the ash accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone, control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively.
[0058] Specifically, in this embodiment, the input parameters of the boiler water wall ash and slagging monitoring model include two-dimensional temperature field and CO concentration distribution, boiler load, coal quality parameters, and flue gas temperature at the furnace outlet, and the output parameter is ash accumulation parameter. The ash accumulation parameter output by the boiler water wall ash and slagging monitoring model is specifically a dynamic characteristic parameter.
[0059] Specifically, in this embodiment, according to the ash accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone, the specific method for controlling the soot blowing operations in the main combustion zone, the burnout zone and the heat exchange zone respectively is: judging whether the dynamic characteristic parameters of each zone are less than the threshold value; if so, it is determined that slagging is occurring in the area, and the soot blower in the corresponding area is controlled to perform soot blowing operations.
[0060] The present invention provides a boiler heating surface slagging monitoring system and method based on a convolutional neural network algorithm. By arranging laser absorption spectrum measurement modules in the main combustion zone, burnout zone and heat exchange zone of the boiler respectively, and then using the adaptive iteration (AART) algorithm, the temperature field and CO concentration field distribution of each part of the heating surface of different sections are first obtained; the artificial convolutional neural network algorithm is adopted to establish a slagging monitoring model for the water-cooled wall of the power station boiler. The established model takes the temperature and CO concentration distribution of different sections, the boiler load, the coal quality, and the flue gas temperature at the furnace outlet as input parameters, and the dynamic characteristic parameters characterizing the pollution status of the furnace heating surface are used as the network output. The slagging state of different sections in the furnace is measured by the dynamic characteristic parameters. Therefore, the present invention can intuitively obtain the slagging state of each heating surface in different sections of the furnace, realize real-time monitoring of the temperatures of different heating surfaces in the furnace, and optimize the over-temperature early warning of the furnace tube and the guidance of the soot blowing frequency.
[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A boiler heating surface slagging monitoring system based on a convolutional neural network algorithm, characterized in that: include: A calculation unit and three laser absorption spectrum measurement modules respectively arranged in the main combustion zone, the burnout zone and the heat exchange zone, wherein the laser absorption spectrum measurement module comprises M first measurement light paths, N second measurement light paths and corresponding detectors, wherein the M first measurement light paths and the N second measurement light paths are arranged vertically and are located in the same horizontal plane, and the M first measurement light paths and the N second measurement light paths divide the measurement area into an M×N grid; the detector is used to measure the transmission spectrum of each measurement light path; The calculation unit includes a two-dimensional field reconstruction unit, a deep learning unit and a soot blowing control unit. The output end of the detector is connected to the two-dimensional field reconstruction unit through an acquisition card. The two-dimensional field reconstruction unit is used to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively according to the spectral information obtained by the detector in each laser absorption spectrum measurement module; The deep learning unit is provided with a boiler water-cooled wall ash and slagging monitoring model established by a convolutional neural network, and the boiler water-cooled wall ash and slagging monitoring model is used to calculate the ash deposition parameters of the main combustion zone, the burnout zone and the heat exchange zone according to the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively; The soot blowing control unit is used to control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively according to the soot accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone.
2. A boiler heating surface slagging monitoring system based on a convolutional neural network algorithm according to claim 1, characterized in that: The input parameters of the boiler water-cooled wall ash and slagging monitoring model include two-dimensional temperature field and CO concentration distribution, boiler load, coal quality parameters, and initial flue gas temperature at the furnace outlet, and the output parameters are dynamic characteristic parameters.
3. A boiler heating surface slagging monitoring system based on a convolutional neural network algorithm according to claim 2, characterized in that: Also includes: A furnace outlet flue gas temperature measuring device, the furnace outlet flue gas temperature measuring device is used to obtain the flue gas temperature at the furnace outlet and send it to the calculation unit; The calculation unit is connected to the electric field DCS system and is used to obtain coal quality parameters and boiler load from the electric field DCS system.
4. A boiler heating surface slagging monitoring system based on a convolutional neural network algorithm according to claim 2, characterized in that: It also includes soot blowers respectively arranged in the main combustion zone, the burnout zone and the heat exchange zone, and the control end of the soot blower is connected to the soot blowing control unit.
5. The boiler heating surface slagging monitoring system based on convolutional neural network algorithm according to claim 1 is characterized in that: The dust accumulation parameter is specifically a dynamic characteristic parameter, and the definition formula of the dynamic characteristic parameter is: Among them, C represents the dynamic characteristic parameters of each region, T a Represents the average furnace temperature in each area, T b ' i Represents the surface temperature distribution value of the slag layer near the wall in the two-dimensional discrete grid of each area.
6. The boiler heating surface slagging monitoring system based on convolutional neural network algorithm according to claim 1 is characterized in that: The specific method of the sootblowing control unit controlling the main combustion zone, the burnout zone and the heat exchange zone to perform sootblowing operations respectively according to the dynamic characteristic parameters of the main combustion zone, the burnout zone and the heat exchange zone is: It is determined whether the dynamic characteristic parameters of each zone are less than a threshold value. If so, it is determined that slagging occurs in the zone, and the soot blower in the corresponding zone is controlled to perform soot blowing operation.
7. The boiler heating surface slagging monitoring system based on convolutional neural network algorithm according to claim 1 is characterized in that: The laser absorption spectrum measurement module also includes a laser controller, a signal generator, a laser, a fiber optic beam splitter and a plurality of collimators. The signal generator is used to input a sawtooth wave signal to the laser controller, and the laser controller is used to control the output laser frequency and scanning range of the laser according to the sawtooth wave signal. The laser output by the laser is divided into M+N beams of laser light after passing through the fiber optic beam splitter, and then passes through a collimator to form a corresponding first measurement optical path and a second measurement optical path.
8. The boiler heating surface slagging monitoring system based on convolutional neural network algorithm according to claim 1 is characterized in that: The two-dimensional field reconstruction unit adopts an adaptive iterative algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion zone, the burnout zone and the heat exchange zone respectively.
9. A boiler heating surface slagging monitoring method based on a convolutional neural network algorithm, characterized in that: The following steps are involved: Step 1: respectively set up laser absorption spectrum measurement modules in the main combustion zone, the burnout zone and the heat exchange zone; Step 2: Obtain the transmission spectrum data of each area, and use the adaptive iterative algorithm to reconstruct the two-dimensional temperature field and CO gas concentration field of the main combustion area, burnout area and heat exchange area respectively; Step 3: Based on the convolutional neural network, a boiler water wall ash and slagging monitoring model is established, and the historical operation data of the power plant is obtained to train the model; Step 4: Input the real-time two-dimensional temperature field and CO gas concentration field data of the main combustion zone, burnout zone and heat exchange zone into the trained model, and output the ash accumulation parameters of the main combustion zone, burnout zone and heat exchange zone respectively through the model; Step 5: According to the ash accumulation parameters of the main combustion zone, the burnout zone and the heat exchange zone, control the main combustion zone, the burnout zone and the heat exchange zone to perform soot blowing operations respectively.
Citation Information
Cited By
Power station boiler temperature field online monitoring method and system based on convolutional neural network and flame radiation image
CN120521730A
Boiler heating surface slagging contamination risk early warning method
CN120974317A
