Boiler multi-parameter intelligent optimization combustion self-adaptive control method for coal blending combustion peak regulation scene

Abstract

The invention relates to the technical field of self-adaptive control, in particular to a boiler multi-parameter intelligent optimization combustion self-adaptive control method for a coal blending combustion peak regulation scene, which comprises the following steps of: acquiring air-coal mixture flow state parameters at an inlet of a combustor in real time, and generating a primary combustion feature set; the primary combustion feature set is processed based on an electric ion radiation tomography algorithm, a three-dimensional thermal kinetic energy distribution map in the combustor is reconstructed, an over fire air compensation coefficient set is output, the three-dimensional thermal kinetic energy distribution map is input into a digital-intelligent twin control platform, and a secondary air door dynamic opening parameter set and a combustor air distribution optimization gradient matrix are generated; and dynamically displaying the optimized gradient matrix through an AR (Augmented Reality) interaction interface, and implementing a combustor throat dynamic compensation strategy based on wall temperature data fed back by an SES-6000 data center cabinet to form a closed-loop optimization loop. According to the method, high-efficiency, low-deviation and wide-adaptation control of boiler combustion in a multi-coal blending combustion scene is realized, and the method has good engineering applicability and intelligent popularization value.

Description

Technical Field

[0001] The present invention relates to the technical field of adaptive control, and particularly to a multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of blended coal types. Background Art

[0002] Currently, when boilers burn multiple coal types and participate in power grid peak shaving operation, their combustion processes present problems such as large coal quality fluctuations, uneven heat load distribution, and high local overheating risks, posing higher requirements for boiler efficiency, safety, and regulation ability; traditional combustion regulation methods based on single-point sensing and static control strategies are difficult to perceive the air-coal mixing state and the internal thermal characteristics of burners in real time, and cannot meet the fine-grained control requirements for the local state of burners. Although existing research has introduced electro-ion sensing and flow field visualization methods for combustion monitoring, there are still limitations such as slow response and weak model adaptability in aspects such as combustion characteristic space analysis, adjustment instruction generation, and feedback execution closed-loop.

[0003] Under this background, there is an urgent need to construct a multi-parameter adaptive combustion control mechanism that integrates multi-source sensing, atlas modeling, digital twin calculation, and augmented reality interaction to achieve a complete closed-loop control process from obtaining the air-coal mixing state, reconstructing the thermal kinetic energy atlas, generating optimized control parameters, to real-time wall temperature feedback and compensation regulation. Summary of the Invention

[0004] The present invention provides a multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of blended coal types, mainly including various parameters such as speed, concentration, flow rate, particle fineness, calorific value, water content, and thermal kinetic energy. It makes full use of the multi-dimensional flow state characteristics under the spatial containment structure, the three-dimensional reconstruction ability based on the electro-ion radiation tomography algorithm, and the AR+ edge fusion platform for control instruction visualization and execution tracking to achieve efficient, safe, and intelligent operation of boiler combustion under complex working conditions.

[0005] A multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of blended coal types includes the following steps:

[0006] S1: Real-time collect the fluid state parameters of the air-coal mixture at the burner inlet through a spatial containment structure sensor array (the fluid state parameters include various parameters such as speed, concentration, flow rate, particle fineness, calorific value, water content, and thermal kinetic energy) to generate a primary combustion feature set including a pulverized coal flow velocity tensor matrix, a concentration gradient cloud map, and a calorific value fluctuation vector. The sensor array includes an MPT-IV type particle impact plate sensor and an electrostatic ion detection module with a collector capacitor plate;

[0007] S2: Process the primary combustion feature set based on the electro-ion radiation tomography algorithm, reconstruct the three-dimensional thermal kinetic energy distribution atlas inside the burner, and output a set of burnout air compensation coefficients;

[0008] S3: Input the three-dimensional thermal kinetic distribution map into the digital twin control platform to generate a set of dynamic opening parameters for the secondary air dampers and an optimization gradient matrix for burner air distribution.

[0009] S4: Dynamically display the optimization gradient matrix through the AR augmented reality interaction interface, and implement a dynamic compensation strategy for the burner throat based on the wall temperature data fed back by the SES-6000 data center cabinet to form a closed-loop optimization circuit.

[0010] Optionally, S1 includes:

[0011] S11: Deploy and initialize the MPT-IV type particle impact plate sensor and the electrostatic ion detection module with a collector capacitor plate to form a spatially inclusive structure sensor array for multi-point distributed monitoring of the air-coal mixture flow channel.

[0012] S12: Based on the multi-point particle impact response model and the electrostatic signal response model, extract the burner inlet characteristics in real time, including the pulverized coal flow velocity tensor matrix, the pulverized coal concentration gradient cloud map, and the calorific value fluctuation vector.

[0013] S13: Combine the above three types of burner inlet characteristics into a primary combustion characteristic set and synchronously transmit it to the subsequent map reconstruction module.

[0014] Optionally, S12 includes:

[0015] S121: Construct the pulverized coal flow velocity tensor matrix according to the impact frequency and particle size information collected by the MPT-IV type particle impact plate sensor.

[0016] S122: Use the electrostatic ion detection module with a collector capacitor plate to obtain multi-point voltage responses, reconstruct the two-dimensional concentration distribution map, and further calculate its spatial gradient to form the pulverized coal concentration gradient cloud map.

[0017] S123: Perform standard deviation analysis on the electrostatic voltage output and impact frequency data at different times, and form the calorific value fluctuation vector through weighted fusion.

[0018] Optionally, S13 includes:

[0019] S131: Normalize the extracted pulverized coal flow velocity tensor matrix, concentration gradient cloud map, and calorific value fluctuation vector respectively to complete the alignment of the unified time axis.

[0020] S132: Feature set construction: Aggregate the normalized features of various types according to the time stamp to construct a multi-dimensional primary combustion characteristic set.

[0021] S133: Feature set transmission: Input it into the electro-ion radiation tomography map reconstruction module in real time through a high-speed data bus.

[0022] Optionally, S2 includes:

[0023] S21, constructing a thermal kinetic projection vector by fusing multi-source features: Align the normalized pulverized coal flow velocity tensor, the pulverized coal concentration gradient cloud map in the multi-dimensional primary combustion feature set, and the normalized voltage response data output by the electrostatic ion detection module in the time dimension, and construct a thermal kinetic projection vector through a weighted fusion method;

[0024] S22, reconstructing a three-dimensional thermal kinetic distribution map based on the fused projection vector: Using the fused thermal kinetic projection vector as the input, perform three-dimensional voxel-level thermal kinetic reconstruction through an electrical ion tomography inversion algorithm, and use a regularization inversion method to solve the underdetermined equation set to generate a thermal kinetic distribution map inside the burner;

[0025] S23, extracting the low thermal energy area and outputting a set of overfire air compensation coefficients: Based on the reconstructed three-dimensional thermal kinetic distribution map, extract the low thermal intensity area near the combustion outlet section, and generate a set of overfire air compensation coefficients through normalized inverse mapping.

[0026] Optionally, S3 includes:

[0027] S31, constructing an air distribution control objective function and partitioning to extract thermal kinetic region features: Input the three-dimensional thermal kinetic distribution map into the digital twin control platform, partition the internal space of the burner, extract the average thermal kinetic level of each partition respectively, and construct an air distribution control function with the thermal kinetic balance degree as the goal, that is, the thermal kinetic balance objective function;

[0028] S32, generating a set of control parameters and outputting the optimized air distribution result: Solve the thermal kinetic balance objective function through the gradient descent optimization algorithm, and output a set of dynamic opening degree parameters for adjusting the secondary air dampers of each partition and the optimized air distribution gradient matrix of the burner.

[0029] Optionally, S31 includes:

[0030] S311, partitioning the thermal kinetic map in space and statistically calculating the average thermal kinetic value in each partition;

[0031] S312, constructing a thermal kinetic balance objective function based on the deviation degree of the thermal kinetic energy of each partition from the global average value.

[0032] Optionally, S32 includes:

[0033] S321, solving the thermal kinetic balance objective function through the gradient descent optimization algorithm, and outputting a set of dynamic opening degree parameters for adjusting the secondary air dampers of each partition;

[0034] S322. Solve the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and output the burner air distribution optimization gradient matrix.

[0035] Optionally, the S4 includes:

[0036] S41. Optimization gradient visualization and interactive instruction synchronization: Input the burner air distribution optimization gradient matrix into the AR augmented reality interactive interface, combine it with the burner three-dimensional structure model for area mapping and color coding, and display the response intensity of each air damper adjustment to the local thermal kinetic energy in real time; at the same time, provide an air damper state adjustment feedback instruction channel to realize the switching between human-machine interaction control and automatic adjustment modes.

[0037] S42. Execution of the throat dynamic compensation strategy based on wall temperature feedback: Real-time obtain the feedback data of the wall-mounted temperature sensors from the SES-6000 data center cabinet, construct the wall temperature fluctuation curve of the throat area, and extract the gradient abnormal points; according to the local overheating or partial cooling trend, adjust the throat air volume distribution factor to form a burner throat dynamic compensation control instruction set, and loop back to the air damper actuator.

[0038] S43. Construct a closed-loop optimization loop and execute feedback control in real time: Jointly input the optimization control parameter set, gradient matrix and throat dynamic compensation factor into the edge controller to generate an air damper execution instruction set, drive the secondary air dampers in each area to achieve dynamic adjustment, and realize the closed-loop control process of data perception - spectrum analysis - instruction optimization - equipment control.

[0039] Optionally, the S42 includes:

[0040] S421. Definition of the wall temperature gradient abnormal coefficient: Construct the wall temperature fluctuation curve of the throat area to identify the areas of thermal unevenness or temperature mutation.

[0041] S422. Calculation of the throat air volume compensation factor: Adjust the throat air volume distribution factor according to the local overheating or partial cooling trend.

[0042] Advantages of the present invention:

[0043] The boiler multi-parameter intelligent optimization combustion adaptive control method for the peak shaving scenario of blended coal types provided by the present invention can real-time obtain the multi-dimensional flow state characteristics of the air-coal mixed flow through the space-inclusive structure sensor array, and combine the electric ion radiation tomography algorithm to accurately reconstruct the three-dimensional distribution map of the internal thermal kinetic energy of the burner, significantly improving the spatio-temporal recognition accuracy of complex combustion states; by fusing multi-source characteristics such as electrostatic signals, pulverized coal flow rate and concentration gradient, it can effectively characterize the fuel mixing uniformity and thermal stability, overcoming the problems of traditional combustion monitoring means relying on a single parameter and local response lag.

[0044] Based on the digital twin control platform, the present invention conducts zonal analysis and optimization solution on the thermal kinetic map, dynamically generates the opening parameter of the secondary air damper and the combustion air distribution gradient matrix of the burner, and presents them visually and conducts human-machine interaction operations through the AR augmented reality interface; at the same time, combined with the wall temperature information fed back by the SES-6000 data center cabinet, a local dynamic compensation strategy for the burner throat area is implemented to form an adaptive closed-loop optimization loop of perception - analysis - adjustment - feedback, realizing high-efficiency, low-deviation, and wide-adaptability control of boiler combustion in the scenario of multi-coal blending, and having good engineering applicability and intelligent popularization value. Brief Description of the Drawings

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only those of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0046] Figure 1 It is the flowchart of the method according to the embodiment of the present invention;

[0047] Figure 2 It is the feature extraction diagram according to the embodiment of the present invention. Detailed Embodiments

[0048] The following will describe the present invention in detail with reference to the drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the drawing part is only for more specific description of the embodiments, and is not intended to specifically limit the present invention.

[0049] It should be pointed out that in the specification, it is mentioned that "an embodiment", "embodiment", "exemplary embodiment", "some embodiments", etc. indicate that the described embodiment may include specific features, structures or characteristics, but not necessarily every embodiment includes such specific features, structures or characteristics. In addition, when combining an embodiment to describe a specific feature, structure or characteristic, implementing such feature, structure or characteristic in combination with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the relevant art.

[0050] Generally, terms can be understood at least in part from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or property in the singular sense, or can be used to describe a combination of features, structures, or properties in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but rather can alternatively, depending at least in part on the context, allow for the existence of other factors that are not necessarily explicitly described.

[0051] As Figure 1 - Figure 2 shown, a multi-parameter intelligent optimization combustion adaptive control method for boilers in the peaking scenario of co-firing coal types includes the following steps:

[0052] S1: Real-time collect the fluid parameters of the air-coal mixture at the burner inlet through a spatially inclusive structure sensor array (the fluid parameters include various parameters such as velocity, concentration, flow rate, particle fineness, calorific value, water content, thermal kinetic energy, etc.), generate a primary combustion feature set including a pulverized coal flow velocity tensor matrix, a concentration gradient cloud map, and a calorific value fluctuation vector, and the sensor array includes an MPT-IV type particle impact plate sensor and an electrostatic ion detection module with a collector capacitor plate;

[0053] S2: Process the primary combustion feature set based on the electrical ion radiation tomography (ES+ETC) algorithm, reconstruct the three-dimensional thermal kinetic energy distribution map inside the burner, and output a set of burnout air compensation coefficients;

[0054] S3: Input the three-dimensional thermal kinetic energy distribution map into the digital twin control platform to generate a set of dynamic opening parameters for the secondary air dampers and an optimization gradient matrix for the burner air distribution;

[0055] S4: Dynamically display the optimization gradient matrix through an AR augmented reality interaction interface, and implement a dynamic compensation strategy for the burner throat based on the wall temperature data fed back by the SES-6000 data center cabinet to form a closed-loop optimization circuit.

[0056] S1 includes:

[0057] S11, deploy and initialize an MPT-IV type particle impact plate sensor and an electrostatic ion detection module with a collector capacitor plate to form a spatially inclusive structure sensor array, and perform multi-point distributed monitoring on the air-coal mixture flow channel;

[0058] S12, based on the multi-point particle impact response model and the electrostatic signal response model, extract the burner inlet characteristics in real time, including a pulverized coal flow velocity tensor matrix, a pulverized coal concentration gradient cloud map, and a calorific value fluctuation vector;

[0059] S13, combine the above three types of burner inlet characteristics into a primary combustion feature set and synchronously transmit it to the subsequent map reconstruction module.

[0060] S12 includes:

[0061] S121, based on the impact frequency and particle size information collected by the MPT-IV particle impact plate sensor, construct the coal powder flow velocity tensor matrix V coal , characterizes the dynamic distribution of fuel particles at multiple locations and in the time dimension, and is expressed as:

[0062]

[0063] in, is the average particle size of the pulverized coal at the i, j position in the sensor array, is the particle impact frequency at the i, j position in the sensor array at time t, obtained by the MPT-IV sensor, A s is the effective sensing area of the particle punch plate sensor;

[0064] S122, using the electrostatic ion detection module with a collector capacitor plate to obtain multi-point voltage responses, reconstruct the two-dimensional concentration distribution map, and further calculate its spatial gradient to form a coal powder concentration gradient cloud map Expressed as:

[0065]

[0066] Where C(x, y) is the instantaneous concentration field of pulverized coal reconstructed by voltage interpolation on the two-dimensional coordinate plane, are the concentration gradient components along the x and y directions;

[0067] S123, perform standard deviation analysis on the electrostatic voltage output and impact frequency data at different times, and form a calorific value fluctuation vector (H) by weighted fusion. var ), which is used to characterize the short-term fluctuation of combustion heat characteristics and is expressed as:

[0068] H var (t)=α·std(U ion (t))+β·std(f c (t));

[0069] Among them, U ion (t) is the output voltage set of all electrostatic ion modules at time t, f c is the set of impact frequencies of all punch plate sensors at time t, std(·) represents the standard deviation operation, α, β are empirically set weighting coefficients used to regulate the impact of voltage and frequency fluctuations on calorific value estimation, α, β∈[0.3, 0.7].

[0070] S13 includes:

[0071] S131. Normalize the obtained pulverized coal flow velocity tensor matrix, concentration gradient nephogram, and calorific value fluctuation vector respectively to eliminate the scale differences between different dimensional features and complete the alignment of the unified time axis, which is expressed as:

[0072]

[0073] Among them, x specifically includes the pulverized coal flow velocity tensor matrix after normalization processing The concentration gradient nephogram after spatial interpolation and standardization processing Normalized calorific value fluctuation vector

[0074] S132. Feature set construction: Aggregate the normalized various features according to the time stamp to construct a multi-dimensional primary combustion feature set It is expressed as:

[0075]

[0076] S133. Feature set transmission: Input it in real time through a high-speed data bus to the electric ion radiation tomography (ES+ETC) map reconstruction module for dynamic deduction of three-dimensional thermal kinetic energy distribution.

[0077] S2 includes:

[0078] S21. Construct a thermal kinetic energy projection vector by fusing multi-source features: Align the normalized pulverized coal flow velocity tensor, pulverized coal concentration gradient nephogram in the multi-dimensional primary combustion feature set with the normalized voltage response data output by the electrostatic ion detection module in the time dimension, and construct a thermal kinetic energy projection vector through a weighted fusion method to comprehensively reflect the spatio-temporal potential characteristics of the thermal kinetic energy distribution inside the burner, which is expressed as:

[0079]

[0080] Among them, P fused (t) is the fused thermal kinetic energy projection vector, is the normalized pulverized coal flow velocity tensor, is the normalized concentration gradient nephogram, is the normalized electric ion response vector, mean(·) is to calculate the mean value of the tensor or image in the spatial dimension, and α1, α2, α3 are feature fusion weight coefficients, satisfying α1 + α2 + α3 = 1, and the value range is [0.2, 0.6];

[0081] S22. Reconstruct the three-dimensional thermal kinetic energy distribution map based on the fused projection vector: Take the fused thermal kinetic energy projection vector as the input, perform three-dimensional voxel-level thermal kinetic energy reconstruction through the electric ion tomography inversion algorithm, and use the regularization inversion method to solve the underdetermined equation set to generate the thermal kinetic energy distribution map inside the burner, which is expressed as:

[0082]

[0083] Among them, W is the electro-ion radiation response matrix, E therm is the three-dimensional thermal kinetic energy voxel vector to be solved, λ is the regularization coefficient, and the value range is 10 -4 ~10 -1 , is the thermal kinetic energy distribution map reconstructed, X, Y, and Z are the spatial dimension resolutions of the map;

[0084] S23, extract the low thermal energy area and output the burnout air compensation coefficient set: Based on the reconstructed three-dimensional thermal kinetic energy distribution map, extract the low thermal intensity area near the combustion outlet section, and generate the burnout air compensation coefficient set through normalized inverse mapping as the input parameter for the subsequent control link to achieve regional precise air supply compensation, expressed as:

[0085]

[0086] Among them, is the burnout air compensation coefficient distribution map, z outlet is the cross-section position near the combustion outlet, γ is the air volume adjustment proportionality coefficient, and the value range is 0.5~1.5, is the maximum thermal kinetic energy value of all voxels in the map for normalization processing.

[0087] S3 includes:

[0088] S31, construct the air distribution control objective function and extract the thermal kinetic energy region characteristics by partition: Input the three-dimensional thermal kinetic energy distribution map into the digital twin control platform, partition the internal space of the burner, extract the average thermal kinetic energy level of each partition respectively, and construct the air distribution control function with the thermal kinetic energy balance degree as the goal, that is, the thermal kinetic energy balance objective function, as the basis for solving the optimization variable;

[0089] S32, generate the control parameter set and output the air distribution optimization result: Solve the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and output the dynamic opening parameter set for adjusting the secondary air dampers of each partition and the air distribution optimization gradient matrix of the burner to drive the actuator to complete the local air volume adaptive adjustment.

[0090] S31 includes:

[0091] S311, perform spatial partitioning on the thermal kinetic energy map, and statistically calculate the average thermal kinetic energy value E i in each partition, expressed as:

[0092]

[0093] Among them, R mis the i-th spatial partition in the thermal kinetic energy map, |R m | is the number of voxels in partition R m , E m is the average thermal kinetic energy of partition m;

[0094] S312 constructs a thermal kinetic energy balance objective function J based on the degree of deviation of the thermal kinetic energy of each partition from the global mean, expressed as:

[0095]

[0096] Among them, is the global average thermal kinetic energy, J is the thermal kinetic energy balance objective function, used as the air distribution optimization target, N is the number of partitions, ranging from 4 to 16.

[0097] S32 includes:

[0098] S321 solves the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and outputs a set of dynamic opening parameters for adjusting the secondary air dampers of each partition, expressed as:

[0099]

[0100] Among them, θ m is the opening control amount of the secondary air damper corresponding to partition m, and the control range is [0, 100], is the initial set opening, η is the learning rate, and the value range is 0.01 to 0.1, is the gradient of the objective function with respect to the opening of the m-th damper;

[0101] S322 solves the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and outputs a burner air distribution optimization gradient matrix, expressed as:

[0102]

[0103] Among them, represents the influence of the change in the opening of the n-th damper on the thermal kinetic energy of the m-th partition, burner air distribution optimization gradient matrix.

[0104] S4 includes:

[0105] S41, optimization gradient visualization and interactive instruction synchronization: Input the burner air distribution optimization gradient matrix into the AR augmented reality interactive interface, combine it with the three-dimensional structure model of the burner for regional mapping and color coding, and display the response intensity of the adjustment of each damper to the local thermal kinetic energy in real time; at the same time, provide a feedback instruction channel for the damper state adjustment to realize the switching between human-computer interaction control and automatic adjustment mode, specifically including:

[0106] (1) In the AR interface, color grading is performed according to the thermal response intensity. For example, red indicates high response and blue indicates weak response;

[0107] (2) The user can select the "automatic execution" or "manual fine-tuning" mode, and the adjustment target is the thermal kinetic energy balance objective function J;

[0108] S42. Execute the dynamic compensation strategy for the throat based on wall temperature feedback: Obtain the feedback data of the temperature sensors arranged on the wall in real time from the SES-6000 data center cabinet, construct the wall temperature fluctuation curve in the throat area, and extract the gradient anomaly points; According to the local overheating or cold bias trend, adjust the air volume distribution factor in the throat to form a dynamic compensation control instruction set for the burner throat, and close the loop and transmit it back to the damper actuator;

[0109] S43. Construct a closed-loop optimization loop and execute feedback control in real time: Jointly input the optimization control parameter set, gradient matrix, and throat dynamic compensation factor into the edge controller to generate a damper execution instruction set, drive the secondary dampers in each area to achieve dynamic adjustment, and realize the closed-loop control process of data perception - spectrum analysis - instruction optimization - equipment control, expressed as:

[0110] Θ exec (X, y) = θ i(x，y) +Δθ throat (x, y);

[0111] Among them, i(x, y) represents the number of the optimization control area to which the coordinate (x, y) belongs, and Θ exec (x, y) is the final damper opening value used to drive the actuator, and θ i(x，y) is the original optimized opening of area i, and Δθ throat (x, y) is the local compensation amount in the throat area;

[0112] If the area x, y is not within the throat range, then Δθ throat (x, y) = 0.

[0113] S42 includes:

[0114] S421. Definition of the wall temperature gradient anomaly coefficient: Construct the wall temperature fluctuation curve in the throat area to identify the area of thermal non-uniformity or temperature mutation, expressed as:

[0115] Among them, T wall (x, y) is the wall temperature distribution in the two-dimensional area of the burner throat, and ξ wall( (x, y) is the wall temperature change rate, used to detect the local temperature rise or sudden drop area;

[0116] S422, Throat air volume compensation factor calculation: Adjust the throat air volume distribution factor according to the local overheating or overcooling trend, expressed as:

[0117] Δθ throat( (x, y) = κ·ξ wall (x, y);

[0118] Among them, Δθ throat( (x, y) is the damper compensation adjustment value corresponding to the throat area, κ is the wall temperature control gain coefficient, and its value range is [0.1, 1.0], which can be set according to the response sensitivity.

[0119] The present invention covers any substitutions, modifications, equivalent methods and solutions made within the spirit and scope of the present invention. For the public to have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention. However, those skilled in the art can also fully understand the present invention without these detailed descriptions. In addition, well-known methods, processes, procedures, components and circuits are not described in detail to avoid unnecessary confusion to the essence of the present invention.

[0120] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of co-firing coal types, characterized in that, It includes the following steps: S1: Real-time collect the fluid state parameters of the air-coal mixture at the burner inlet through a spatial containment structure sensor array, and generate a primary combustion feature set including a pulverized coal flow velocity tensor matrix, a concentration gradient cloud map, and a calorific value fluctuation vector. The sensor array includes an MPT-IV type particle impact plate sensor and an electrostatic ion detection module with a collector capacitor plate; S2: Process the primary combustion feature set based on the electro-ion radiation tomography algorithm, reconstruct the three-dimensional thermal kinetic energy distribution map inside the burner, and output a set of burnout air compensation coefficients; S3: Input the three-dimensional thermal kinetic energy distribution map into the digital twin control platform to generate a set of dynamic opening parameters for the secondary air dampers and an optimization gradient matrix for the burner air distribution; S4: Dynamically display the optimization gradient matrix through an AR augmented reality interaction interface, and implement a dynamic compensation strategy for the burner throat based on the wall temperature data fed back by the SES-6000 data center cabinet to form a closed-loop optimization circuit.

2. The multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario facing blended coal types according to claim 1, wherein, The S1 includes: S11. Deploy and initialize the MPT-IV type particle impact plate sensor and the electrostatic ion detection module with a collector capacitor plate to form a spatial containment structure sensor array, and perform multi-point distributed monitoring on the air-coal mixture flow channel; S12. Based on the multi-point particle impact response model and the electrostatic signal response model, real-time extract the burner inlet characteristics, including the pulverized coal flow velocity tensor matrix, the pulverized coal concentration gradient cloud map, and the calorific value fluctuation vector; S13. Combine the above three types of burner inlet characteristics into a primary combustion feature set and synchronously transmit it to the subsequent map reconstruction module.

3. The multi-parameter intelligent optimized combustion adaptive control method for boilers in a peak shaving scenario for co-firing coal types according to claim 2, wherein, The S12 includes: S121. Construct a pulverized coal flow velocity tensor matrix according to the impact frequency and particle size information collected by the MPT-IV type particle impact plate sensor; S122. Use the electrostatic ion detection module with a collector capacitor plate to obtain the multi-point voltage response, reconstruct the two-dimensional concentration distribution map, and further calculate its spatial gradient to form a pulverized coal concentration gradient cloud map; S123. Perform standard deviation analysis on the electrostatic voltage output and impact frequency data at different times, and form a calorific value fluctuation vector through a weighted fusion method.

4. The multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of blended coal types according to claim 2, wherein The S13 includes: S131. Respectively perform normalization processing on the extracted pulverized coal flow velocity tensor matrix, concentration gradient cloud map, and calorific value fluctuation vector to complete the alignment of the unified time axis; S132. Feature set construction: Aggregate the normalized features of each type according to the time stamp to construct a multi-dimensional primary combustion feature set; S133. Feature set transmission: Input it into the electro-ion radiation tomography map reconstruction module in real time through a high-speed data bus.

5. The intelligent optimization combustion adaptive control method for multi-parameters of a boiler in a peak shaving scenario for co-firing coal types according to claim 4, wherein The S2 includes: S21. Fusion of multi-source features to construct a thermal kinetic energy projection vector: Align the normalized pulverized coal flow velocity tensor, pulverized coal concentration gradient cloud map in the multi-dimensional primary combustion feature set with the normalized voltage response data output by the electrostatic ion detection module in the time dimension, and construct a thermal kinetic energy projection vector through a weighted fusion method; S22. Reconstruct the three-dimensional thermal kinetic energy distribution map based on the fused projection vector: Using the fused thermal kinetic energy projection vector as the input, perform three-dimensional voxel-level thermal kinetic energy reconstruction through the electrical ion tomography inversion algorithm, and use the regularization inversion method to solve the underdetermined equation set to generate the thermal kinetic energy distribution map inside the burner; S23. Extract the low thermal energy region and output the overfire air compensation coefficient set: Based on the reconstructed three-dimensional thermal kinetic energy distribution map, extract the low thermal intensity region near the combustion outlet section, and generate the overfire air compensation coefficient set through normalized inverse mapping.

6. The multi-parameter intelligent optimized combustion adaptive control method for boilers in the peaking scenario of co-firing coal types according to claim 5, characterized in that, The S3 includes: S31. Construct the air distribution control objective function and partition to extract the thermal kinetic energy region characteristics: Input the three-dimensional thermal kinetic energy distribution map into the digital twin control platform, partition the internal space of the burner, extract the average thermal kinetic energy level of each partition respectively, and construct the air distribution control function with the thermal kinetic energy balance as the goal, that is, the thermal kinetic energy balance objective function; S32. Generate the control parameter set and output the air distribution optimization result: Solve the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and output the dynamic opening degree parameter set for adjusting the secondary air dampers of each partition and the air distribution optimization gradient matrix of the burner.

7. A multi-parameter intelligent optimized combustion adaptive control method for boilers in a peak shaving scenario for co-firing coal types according to claim 6, characterized in that The S31 includes: S311. Partition the thermal kinetic energy map in space and statistically calculate the average thermal kinetic energy value in each partition; S312. Based on the deviation degree of the thermal kinetic energy of each partition from the global average value, construct the thermal kinetic energy balance objective function.

8. A multi-parameter intelligent optimized combustion adaptive control method for boilers in a peaking scenario facing blended coal types according to claim 6, characterized in that, The S32 includes: S321. Solve the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and output the dynamic opening degree parameter set for adjusting the secondary air dampers of each partition; S322. Solve the thermal kinetic energy balance objective function through the gradient descent optimization algorithm, and output the air distribution optimization gradient matrix of the burner.

9. The multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of blended coal types according to claim 8, wherein, The S4 includes: S41. Optimize gradient visualization and synchronize interaction instructions: Input the air distribution optimization gradient matrix of the burner into the AR augmented reality interaction interface, combine with the three-dimensional structure model of the burner for regional mapping and color coding, and display the response intensity of each damper adjustment to local thermal kinetic energy in real time; at the same time, provide a feedback instruction channel for damper state adjustment to realize the switching between human-computer interaction control and automatic adjustment mode; S42. Execute the dynamic compensation strategy for the throat based on wall temperature feedback: Obtain the feedback data of the wall-mounted temperature sensors in real time from the SES-6000 data center cabinet, construct the wall temperature fluctuation curve of the throat area, and extract the gradient abnormal points; according to the local overheating or cooling trend, adjust the throat air volume distribution factor to form the dynamic compensation control instruction set for the burner throat, and close-loop back to the damper actuator; S43. Construct a closed-loop optimization loop and execute feedback control in real time: Jointly input the optimized control parameter set, gradient matrix and throat dynamic compensation factor into the edge controller to generate the damper execution instruction set, drive the secondary air dampers of each region to achieve dynamic adjustment, and realize the closed-loop control process of data perception - map analysis - instruction optimization - equipment control.

10. A multi-parameter intelligent optimized combustion adaptive control method for boilers in the peak shaving scenario of co-firing coal types according to claim 9, characterized in that The S42 includes: S421. Define the wall temperature gradient abnormal coefficient: Construct the wall temperature fluctuation curve of the throat area to identify the regions of thermal non-uniformity or temperature mutation; S422, Throat air volume compensation factor calculation: Adjust the throat air volume distribution factor according to the local overheating or overcooling trend.

Images