Method and system for closed-loop regulation of thermal performance of a kiln burner

Abstract

The present application relates to the technical field of kiln head burner control, in particular to a thermal performance closed-loop regulation method and system for a kiln head burner. The method comprises: obtaining burner monitoring parameters; converting the monitoring parameters into core performance indexes including primary air momentum, thrust rate, primary air proportion, swirl ratio and kiln skin characteristic parameters in real time based on fluid mechanics and thermodynamics principles; inputting the core performance indexes into an index analysis model to output a comprehensive score of the burner based on the core performance indexes and a real-time regulation strategy based on the deviation of each core performance index from the corresponding preset ideal threshold; regulating the kiln head burner based on the real-time regulation strategy and performing causal correlation analysis on the core performance indexes after regulation to iteratively optimize the fusion strategy of the index analysis model. The system breaks through the bottleneck of the regulation mode of the kiln head burner which relies on implicit experience and provides key support for the digitalization and intelligent control of the cement clinker burning process.

Description

Technical Field

[0001] This invention relates to the field of kiln head burner control technology, specifically to a closed-loop control method and system for the thermal performance of kiln head burners. Background Technology

[0002] In the cement clinker calcination process, the calcination conditions inside the rotary kiln directly determine the quality of the clinker and energy consumption, while the performance and adjustment of the kiln head burner are the core means of controlling the calcination temperature field and flame shape.

[0003] Currently, the industry commonly uses multi-duct pulverized coal burners as kiln head burners. These typically rely on one or more fans (such as Roots blowers or magnetic levitation blowers) to provide primary air. The air duct structure is generally divided into external air, coal air, internal air, and central air. In actual production operation, the air volume and flame shape are usually adjusted by regulating the opening of the butterfly valves for the external, internal, and central air, or by changing the cross-sectional area of ​​the combustion nozzle. However, the fundamental contradiction between the "black box" state of the thermal performance of existing kiln head burners and the demands for intelligent and refined production lies in the fact that current technology treats the burner as an open-loop control object relying on "input-observation-experience adjustment," which presents the following technical problems:

[0004] 1. The gas-solid two-phase flow, combustion, and heat transfer processes inside the kiln head burner are extremely complex. Key performance parameters (such as actual thermal thrust and swirl intensity) cannot be directly, online, and continuously obtained using precision instruments. This leads to our understanding of the kiln head burner's operating status remaining at the level of indirect parameters like inlet air pressure and air volume, and macroscopic phenomena such as outlet flame and kiln conditions. We fail to grasp the essential fluid dynamics and thermodynamic parameters that determine flame morphology and combustion efficiency. The control process of the kiln head burner is thus driven by "sensory experience" rather than "data mechanism."

[0005] 2. Due to a lack of quantitative understanding of key performance parameters, there is a lack of unified and objective comprehensive evaluation standards for the current operating level of kiln head burners. Therefore, the operation and adjustment of existing kiln head burners rely on the operator's one-sided interpretation of multiple isolated process parameters (such as cylinder scanning temperature, flue gas composition, etc.) and personal experience. The direction of adjustment (such as whether to increase or decrease airflow) and the magnitude of adjustment (how much airflow to increase) lack quantitative basis, resulting in a highly blind adjustment process, long time consumption, high trial and error costs, poor consistency of clinker firing quality, and difficulty in solidifying excellent operating experience.

[0006] 3. There is a lack of real-time, precise correlation mapping and closed-loop verification between the adjustment actions of the kiln head burner and the final process effect. The impact of a single, time-consuming adjustment on the kiln lining condition and pulverized coal burnout rate is severely delayed and subject to multiple interferences such as frequent fluctuations in the kiln's internal operating conditions. Therefore, existing technologies cannot establish a rapid feedback optimization loop between burner operation adjustments and calcination effects, causing burner optimization to remain at a static, local level, and failing to achieve dynamic optimization based on a globally optimal goal.

[0007] Therefore, existing technologies urgently need to study how to transform the cement rotary kiln burner, a thermal device that relies on implicit experience-based adjustment, into an explicit intelligent object with transparent and calculable core parameters, real-time quantifiable operating status, evidence-based adjustment operations, and continuous closed-loop optimization based on effect feedback. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention proposes a closed-loop control method and system for the thermal performance of kiln head burners. This aims to systematically break through the bottleneck of the kiln head burner's adjustment mode, which relies on implicit experience, and provide key support for the digital and intelligent management and control of the cement clinker calcination process.

[0009] Firstly, this application provides a closed-loop control method for the thermal performance of a kiln head burner, comprising the following steps:

[0010] Acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared images, and fuel parameters;

[0011] Based on the principles of fluid mechanics and thermodynamics, the monitoring parameter data is converted into core performance indicators in real time. These core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters.

[0012] A performance index analysis model is constructed. Core performance indexes are input into the performance index analysis model. The performance index analysis model outputs a comprehensive burner score based on the fusion of core performance indexes and a real-time control strategy based on the analysis of the deviation of each core performance index from the corresponding preset ideal threshold.

[0013] The kiln head burner is controlled based on a real-time control strategy, and the core performance indicators after control are continuously collected.

[0014] Causal correlation analysis was conducted on the core performance indicators after regulation, and the fusion strategy of the indicator analysis model was iteratively optimized based on the results of the causal correlation analysis.

[0015] In some embodiments, the fuel parameters include kiln head coal consumption and lower heating value of pulverized coal. Based on fluid dynamics and thermodynamics principles, the monitoring parameter data are converted into core performance indicators in real time, including:

[0016] The kiln heat load is obtained by multiplying the kiln head coal consumption by the lower heating value of pulverized coal and dividing by the unit clinker output.

[0017] The ratio of primary air velocity to kiln heat load is calculated to obtain the thrust rate, which is used to measure the energy utilization efficiency of the burner.

[0018] In some embodiments, the airflow parameter includes the rotary air mass flow rate. Based on fluid dynamics and thermodynamics principles, the monitored parameter data is converted into core performance indicators in real time, including:

[0019] The tangential velocity component of the airflow is calculated based on the principles of fluid mechanics.

[0020] The radial impulse is obtained by calculating the product of the rotating air mass flow rate and the tangential velocity component of the airflow.

[0021] Obtain the burner structural parameters, including the equivalent radius of the swirl duct;

[0022] The product of radial impulse and the equivalent radius of the swirl channel is calculated to obtain the rotational torque of the airflow in the tangential direction;

[0023] Calculate the sum of the primary wind momentum of each axial duct to obtain the total axial impulse;

[0024] The equivalent diameter of the pulse is calculated based on the mass flow rate of the rotating air and the total axial impulse.

[0025] The axial moment of the airflow in the axial direction is obtained by calculating the product of the total axial impulse and the equivalent diameter of the pulse.

[0026] The ratio of the rotational torque of the airflow in the tangential direction to the axial torque in the axial direction is calculated to obtain the swirl ratio, which is used to quantify the rotational intensity of the airflow at the burner outlet.

[0027] In some embodiments, the air volume parameter includes primary air volume, and the monitoring parameter data is converted into core performance indicators in real time based on fluid mechanics and thermodynamics principles, including:

[0028] Based on fuel parameters, the actual total combustion air volume, including primary air volume and high-temperature secondary air volume, is calculated using the solid fuel theoretical air algorithm.

[0029] The ratio of primary air volume to actual total combustion air volume is calculated and converted into a percentage to obtain the primary air proportion, which is used to quantify the balance between the combustion process's dependence on primary air volume and the utilization efficiency of high-temperature secondary air.

[0030] In some embodiments, monitoring parameter data is converted into core performance indicators in real time based on fluid mechanics and thermodynamics principles, including:

[0031] Temperature calibration was performed on the infrared image of the kiln shell to obtain the first temperature matrix;

[0032] The temperature values ​​along the circumference of the cylinder in the first temperature matrix are averaged and reduced in dimension to obtain the second temperature matrix.

[0033] The pixel coordinates along the kiln length in the second temperature matrix are mapped to physical dimensions to obtain the third temperature matrix;

[0034] The length of the kiln lining coverage area was obtained by calculating the third temperature matrix using the RSI trend analysis algorithm.

[0035] The kiln lining thickness characterization is determined based on the relative temperature change of the kiln lining coverage area in the third temperature matrix.

[0036] The kiln skin flatness is calculated based on the temperature values ​​of the kiln skin coverage area in the third temperature matrix along the circumference of the cylinder and the length of the kiln.

[0037] The length of the kiln lining coverage area, the thickness of the kiln lining, and the flatness of the kiln lining constitute the kiln lining characteristic parameters.

[0038] In some embodiments, the RSI trend analysis algorithm is used to calculate the length of the kiln lining coverage area, including:

[0039] The temperature values ​​of the kiln-length direction in the kiln-covered area in the third temperature matrix are denoted as a one-dimensional temperature sequence.

[0040] The one-dimensional temperature sequence is divided according to a preset short period to obtain a short-period moving average sequence.

[0041] The one-dimensional temperature sequence is divided according to a preset long period to obtain a long-period moving average sequence.

[0042] According to the temperature values ​​from high to low, the short-period moving average sequence and the long-period moving average sequence are scanned respectively, and the cross condition judgment is performed during the scanning process to determine the end of the kiln skin.

[0043] The length of the area covered by the kiln lining is calculated based on the coordinate positioning of the end of the kiln lining.

[0044] In some embodiments, the kiln lining smoothness is calculated based on the temperature values ​​of the kiln lining covered area in the third temperature matrix along the circumference of the kiln body and the length of the kiln, including:

[0045] The mean of the temperature standard deviation of the kiln lining covered area in the circumferential direction of the cylinder in the third temperature matrix is ​​calculated to obtain the longitudinal flatness of the matrix.

[0046] Calculate the temperature fluctuation frequency of the kiln lining covered area along the kiln length in the third temperature matrix to obtain the matrix lateral flatness.

[0047] The kiln skin flatness is obtained by weighted fusion of the matrix longitudinal flatness and matrix transverse flatness.

[0048] In some embodiments, the monitoring parameter data further includes smoke chamber process parameters, and the index analysis model outputs a comprehensive burner score based on the fusion of core performance indicators, including:

[0049] The core performance indicators and smoke chamber process parameters are used as evaluation indicators, and scoring functions and corresponding initial weights are set for the evaluation indicators.

[0050] The current values ​​of each evaluation indicator are calculated based on the scoring function, and then standardized to obtain the current score value of each evaluation indicator.

[0051] The current scores of each evaluation indicator and their corresponding initial weights are weighted and summed to obtain the overall score of the burner.

[0052] In some embodiments, causal correlation analysis is performed on the core performance indicators after regulation, and the fusion strategy of the indicator analysis model is iteratively optimized based on the causal correlation analysis results, including:

[0053] The primary air momentum, thrust rate, primary air ratio, and swirl ratio after adjustment were used as dependent variables, and the kiln skin characteristic parameters and smoke chamber process parameters after adjustment were used as resultant variables.

[0054] Time alignment is performed on the dependent and result variables within the preset monitoring period, and the real-time control strategy, control timestamp, dependent and result variables are combined to form a data sample;

[0055] Feature extraction is performed on the data samples to obtain the input features and target variables;

[0056] Construct a machine learning model, train the machine learning model based on input features and target variables, and output the machine learning model after training as the influence coefficients of each dependent variable on the outcome variable;

[0057] Based on the influence coefficients of each dependent variable on the result variable, the initial weights corresponding to each evaluation index are optimized, and the optimized weights are used in the next fusion process of burner comprehensive scoring.

[0058] Secondly, this application provides a closed-loop control system for the thermal performance of a kiln head burner, including:

[0059] The data monitoring module is used to acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared images, and fuel parameters.

[0060] The index calculation module is used to convert monitoring parameter data into core performance indicators in real time based on the principles of fluid mechanics and thermodynamics. The core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters.

[0061] The comprehensive analysis module is used to build an indicator analysis model. The core performance indicators are input into the indicator analysis model, and the indicator analysis model outputs a comprehensive score of the burner based on the fusion of the core performance indicators, as well as a real-time control strategy based on the analysis of the deviation of each core performance indicator from the corresponding preset ideal threshold.

[0062] The control module is used to control the kiln head burner based on a real-time control strategy and continuously collect the core performance indicators after control.

[0063] The feedback optimization module is used to perform causal correlation analysis on the core performance indicators after regulation, and to iteratively optimize the fusion strategy of the indicator analysis model based on the causal correlation analysis results.

[0064] The beneficial technical effects of the present invention include at least the following:

[0065] 1. By constructing a collaborative technology system of "transparent monitoring parameters, quantifiable burner status, evidence-based regulation, and closed-loop adaptive control," this application fundamentally breaks through the bottleneck of the kiln head burner's reliance on implicit experience for regulation. Its core lies in transforming the burner from a "black box" experience-based device into a "white box" intelligent control object, providing quantitative decision support for the digitalization and intelligentization of the clinker firing process. Specifically, this application first achieves a breakthrough at the data perception level. Based on the principles of fluid mechanics and thermodynamics, easily measurable parameters such as air volume parameters, kiln shell infrared images, and fuel parameters are calculated in real time into core performance indicators reflecting the essence of combustion fluid dynamics, such as primary air momentum, thrust rate, primary air ratio, and swirl ratio. This is equivalent to installing a "mechanism vision" on the burner that can see through the internal flow and combustion state, upgrading the previous vague understanding based only on macroscopic phenomena such as flame morphology to a precise quantitative understanding of key performance parameters such as thermal thrust and swirl intensity, solving the fundamental technical problem of the burner's invisible and unmeasurable status. Next, based on this, an index analysis model was constructed. This model provides an objective and unified comprehensive score through the fusion of multiple indicators, and directly generates real-time control strategies based on the deviations of each indicator from the ideal threshold. This liberates the control of the kiln head burner from the predicament of relying on personal experience and one-sided interpretation of multiple parameters, ensuring that each adjustment of the kiln head burner has a clear quantitative target and basis, achieving a leap from "experience-based trial and error" to "model optimization." More importantly, steps S4 and S5 constitute a complete "analysis-suggestion-control-optimization" closed loop. After the system executes the real-time control strategy, it continuously collects core performance indicators and performs causal correlation analysis to iteratively optimize the model's fusion and decision-making logic. This allows the system to dynamically adapt to changes in operating conditions, continuously approaching the global optimum, and ultimately forming an intelligent control loop with self-learning and adaptive capabilities. Therefore, this application achieves state transparency through mechanism transformation, quantifies decision-making through model analysis, and continuously optimizes through closed-loop iteration, thus collaboratively constructing the cornerstone of digital control of kiln head burners. This transforms burner adjustment from an experience-based engineering process into a calculable, optimizable, and solidifiable precision engineering process, providing a key breakthrough and a reliable data-driven paradigm for the intelligent upgrading of the entire cement firing system.

[0066] 2. In traditional cement firing operations, focusing solely on momentum is insufficient. Operators face a fundamental contradiction in their understanding and decision-making regarding kiln head burner adjustments. On one hand, operators know through experience that a sufficiently large primary air momentum is needed to obtain a rigid and penetrating flame, ensuring stable mixing of pulverized coal and air, and combustion. On the other hand, operators also understand that primary air is cold air at a lower temperature, and excessive use (usually reflected in primary air volume) will lower flame temperature, increase clinker heat consumption, and potentially increase NOx formation. However, in traditional understanding and operation, "primary air momentum" (mechanical properties) and "coal consumption" (thermal performance) are two independent parameters monitored and considered separately. Operators cannot intuitively determine whether the current air volume / momentum configuration is optimal in terms of energy utilization efficiency. In response, this application no longer treats "momentum" and "air volume / heat consumption" as two independent variables that require manual trade-offs. Instead, it constructs "thrust rate," a core performance indicator that can directly measure the energy utilization efficiency of the burner, and correlates and normalizes the fluid dynamic performance (thrust) of the burner with its thermal energy consumption performance (heat load). This achieves a shift from "empirical balance of discrete parameters" to "quantitative optimization of coupled parameters." Through a simple ratio, it achieves deep integration and unification, providing a clear quantitative path to resolve long-standing operational contradictions.

[0067] 3. In traditional cement firing operations, operators typically only indirectly perceive primary air volume through fan current and valve opening, lacking effective direct measurement methods for high-temperature secondary air volume. They can usually only vaguely infer it from parameters such as kiln head negative pressure and grate cooler pressure. Furthermore, when adjusting the valve opening for primary air volume, the goal is often to stabilize the flame shape or kiln head temperature, making it impossible to determine whether it is within the optimal range relative to the current coal quality and heat load. Operators theoretically know that using more high-temperature secondary air can save energy and reduce nitrogen, but due to concerns about flame instability and incomplete coal combustion, operations tend to be conservative, hesitant to reduce the primary air volume to the theoretical lower limit, thus failing to fully tap the energy-saving potential. This contradiction stems from the inability to quantitatively assess "how much the primary air volume can be reduced to at least while ensuring sufficient jet flow." In this regard, this application no longer treats primary air as an independent operational variable. Instead, it fills the gap in real-time data of high-temperature secondary air volume by integrating theoretical modeling with online data. It directly quantifies the share of cost air (i.e., primary air volume) in the total air volume. Reducing the proportion of primary air means that, under the premise of meeting the total oxygen demand, the enthalpy of high-temperature secondary air is maximized, reducing the heat required to heat room-temperature primary air, thereby directly reducing clinker heat consumption, and inhibiting the formation of thermal NOx due to the possible reduction in flame peak temperature.

[0068] 4. In existing cement firing operations, operators adjust the swirl and axial flow valves to change the flame shape. However, the core fluid dynamics parameter determining the flame shape, "swirl intensity," cannot be directly observed or measured. Operators can only infer the appropriateness of the swirl intensity based on macroscopic phenomena such as the "short and thick" or "long and thin" shape of the flame inside the kiln, combined with lagging parameters such as kiln tail temperature and kiln lining condition. This "seeing the effect and adjusting the cause" approach in existing technologies suffers from severe lag and blindness. Therefore, this application no longer relies on subjective judgment of the flame shape but directly calculates and monitors the root cause of its formation. By measuring the "disturbance" ability of the rotating air to the axial air, it achieves, for the first time, the quantification of the rotation intensity of the burner outlet airflow. This provides operators with precise numerical targets for adjusting the swirl / axial flow valves, thereby fundamentally solving the problems of severe lag and blindness in existing adjustment technologies.

[0069] Other features and advantages of the present invention will be disclosed in detail in the following detailed description and accompanying drawings. Attached Figure Description

[0070] The invention will be further described below with reference to the accompanying drawings:

[0071] Figure 1 This is a flowchart of the closed-loop control method for the thermal performance of the kiln head burner in an embodiment of the present invention.

[0072] Figure 2 This is a first visual panel schematic diagram of the thermal performance of the kiln head burner in an embodiment of the present invention.

[0073] Figure 3 This is a schematic diagram of the second visualization panel for the thermal performance of the kiln head burner in an embodiment of the present invention.

[0074] Figure 4 This is a schematic diagram of the closed-loop control system for the thermal performance of the kiln head burner in an embodiment of the present invention. Detailed Implementation

[0075] The technical solutions of the embodiments of the present invention will be explained and described below with reference to the accompanying drawings. However, the following embodiments are only preferred embodiments of the present invention and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments in the implementation methods without creative effort are all within the protection scope of the present invention.

[0076] In the following description, terms such as “inner,” “outer,” “upper,” “lower,” “left,” and “right” are used only to indicate orientation or positional relationship for the convenience of describing the embodiments and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0077] Example 1:

[0078] Please see the appendix Figure 1 , Figure 1 A schematic flowchart of a closed-loop control method for the thermal performance of a kiln head burner provided in one embodiment of this specification is shown.

[0079] like Figure 1 As shown, the closed-loop control method for the thermal performance of the kiln head burner may include at least the following steps:

[0080] S1, acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared image and fuel parameters.

[0081] The monitoring parameter data may include at least:

[0082] 1. Airflow parameters: including rotary air mass flow rate (kg / s), primary air mass flow rate (kg / s), and primary air volume (Nm³) for each air duct (external air, coal air, internal air, and central air). 3 / h). Mass flow rate can be measured in various ways. The most common method is using a thermal gas mass flow meter, which can directly measure the mass flow rate without temperature and pressure compensation. Alternatively, differential pressure can be measured using a differential pressure flow meter (such as a venturi tube or airfoil type), and the mass flow rate can be calculated by combining the temperature and pressure data. Primary air volume is the air volume data directly measured by hardware. It differs from "primary air momentum." Primary air volume is the instantaneous primary air volume that is directly and continuously collected and transmitted by online flow meters (e.g., flow meters built into or added to magnetic levitation or air-levitation fans) installed in the primary air main duct or various air ducts.

[0083] 2. Air pressure parameters: This includes installing pressure transmitters in each air duct of the burner to collect real-time data on the external air, coal air, internal air, and central air pressure (kPa).

[0084] 3. Smoke chamber process parameters: including kiln tail smoke chamber temperature (°C), smoke chamber oxygen concentration, and smoke chamber carbon monoxide concentration (ppm) read from DCS;

[0085] 4. Infrared image of the kiln shell; automatically acquired image of the temperature distribution on the surface of the kiln shell, used to analyze the condition of the kiln lining;

[0086] 5. Fuel parameters: including the kiln head coal consumption (kg / s) and the lower heating value of coal (kJ / kg).

[0087] S2, based on the principles of fluid mechanics and thermodynamics, transforms monitoring parameter data into core performance indicators in real time. These core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters.

[0088] In this embodiment, the core performance indicators are not simply data displays, but rather, by using the principles of fluid mechanics and thermodynamics, easily measurable monitoring parameters are transformed into essential parameters (i.e., core performance indicators) that are difficult to measure directly and reflect the core working mechanism of the burner, thus achieving an upgrade in the cognitive dimension.

[0089] Primary air momentum measures the mechanical impact force of the gas ejected from the burner. It is a core fluid dynamics parameter that determines the flame shape, rigidity, and penetration power. The magnitude of momentum directly affects whether the burner can effectively entrain high-temperature secondary air. Excessive momentum can lead to short, thick, and dispersed flames, even damaging the kiln lining. Insufficient momentum results in long but weak flames lacking penetration power, failing to meet the temperature distribution required for clinker calcination. The calculation of primary air momentum is a current technology, and its specific implementation method is as follows:

[0090] S211. Obtain the burner structural parameters, mainly including the nozzle cross-sectional area (unit: square meters) of each air duct such as external air, coal air, internal air, and central air. These are the basic geometric parameters for calculating wind speed;

[0091] S212. Based on the nozzle cross-sectional area of ​​each air duct and the primary air volume of each air duct, the average outlet velocity of the primary air is calculated (unit: m / s).

[0092] S213. Primary air momentum (unit: Newton, N) = primary air mass flow rate × primary air outlet average velocity.

[0093] Thrust ratio measures the burner's energy efficiency, i.e., generating sufficient thrust with as little primary air as possible. Primary air ratio quantifies the balance between the combustion process's dependence on primary air volume and the efficiency of utilizing high-temperature secondary air. Swirl ratio quantifies the intensity of the airflow rotation at the burner outlet; it determines the flame shape (short and thick or long and thin) and the intensity of high-temperature flue gas recirculation, affecting the mixing of pulverized coal and air. Kiln lining characteristic parameters include kiln lining length, kiln lining thickness, and kiln lining smoothness.

[0094] Understandably, focusing solely on momentum is insufficient in existing cement firing operations. Operators face a fundamental contradiction in their understanding and decision-making regarding kiln burner adjustments. On one hand, operators know through experience that a sufficiently large primary air momentum is needed to obtain a rigid and penetrating flame, ensuring stable mixing and combustion of pulverized coal and air. On the other hand, operators also understand that primary air is cold air at a lower temperature, and excessive usage (usually reflected in primary air volume) will lower flame temperature, increase clinker heat consumption, and potentially increase NOx formation. However, in current understanding and operations, "primary air momentum" (mechanical performance) and "coal consumption" (thermal performance) are two independent parameters monitored and considered separately. Operators cannot intuitively determine whether the current air volume / momentum configuration is optimal in terms of energy utilization efficiency. Existing technology lacks a unified, quantifiable, and balanced core performance indicator that addresses this contradiction. Therefore, this embodiment creatively introduces and calculates in real-time the comprehensive evaluation indicator of "thrust rate." Thrust ratio reflects whether a burner can generate a powerful jet impact force with extremely high wind speed using as little primary air as possible, thereby meeting the aerodynamic requirements of fuel combustion.

[0095] Specifically, based on the principles of fluid mechanics and thermodynamics, the monitored parameter data is converted into thrust rate in real time. The specific implementation method is as follows:

[0096] S221, Kiln heat load (unit: MW) = Kiln head coal powder consumption (kg / s) × coal powder lower heating value (kJ / kg) / 1000;

[0097] S222, Thrust rate (unit: N / MW) = Primary air velocity / Kiln heat load.

[0098] It can be seen that the difference between thrust ratio and primary air momentum is that primary air momentum focuses on fluid dynamics characteristics, which directly determines the penetration force, range, and ability to entrain high-temperature secondary air in the furnace; while thrust ratio focuses on thermal and energy consumption characteristics, which measures the energy utilization efficiency of the burner. The higher the thrust ratio, the more the burner can generate higher momentum with less ambient temperature primary air, thus making more effective use of the high-temperature secondary air from the grate cooler.

[0099] Therefore, the technical concept of this solution lies in correlating and normalizing the hydrodynamic performance (thrust) of the burner with its thermal energy consumption performance (heat load), creating a core performance indicator that can directly measure the energy utilization efficiency of the burner. The core idea is that evaluating the quality of a burner should not only consider the amount of thrust it generates, but also how much mechanical thrust benefit it generates for every unit of thermal energy cost. This aims to guide the burner's control operation from pursuing the extreme value of a single parameter to seeking the optimal energy efficiency ratio of the entire system.

[0100] Understandably, the improved approach of this solution represents a shift from "empirical balancing of discrete parameters" to "quantitative optimization of coupled parameters." This embodiment no longer treats "momentum" and "airflow / heat consumption" as two independent variables requiring manual trade-offs. Instead, it internalizes their relationship into a directly monitorable and optimizeable target value by constructing a new core performance indicator: "thrust rate." This simplifies the burner's control operation from a vague dual objective of achieving momentum targets while minimizing primary airflow to a single, explicit objective of stabilizing the thrust rate within the optimal range.

[0101] It is understandable that the "thrust rate" introduced in this embodiment is not a simple formula calculation, but a key innovation in the burner control concept. The thrust rate deeply integrates and unifies the originally opposing and separate evaluation dimensions of "mechanics" and "thermality" through a concise ratio, providing a clear quantitative path to resolve long-standing operational contradictions. By monitoring and optimizing this core performance indicator of thrust rate in real time, the system can automatically find its way to the high-efficiency operating range. This is one of the most direct and core technical mechanisms for reducing energy consumption and improving thermal efficiency.

[0102] In existing cement firing operations, operators' understanding and control of "primary air volume" and "secondary air volume" are disconnected. They can usually only indirectly perceive the primary air volume through fan current and valve opening, while lacking effective direct measurement methods for high-temperature secondary air volume. They can only vaguely infer it from parameters such as kiln head negative pressure and grate cooler pressure, resulting in a one-sided perception. Moreover, when adjusting the valve opening for primary air volume, the goal is often to stabilize the flame shape or kiln head temperature, without being able to determine whether it is within the optimal range relative to the current coal quality and heat load. Operators theoretically know that using more high-temperature secondary air can save energy and reduce nitrogen, but due to concerns about flame instability and incomplete coal combustion, operations tend to be conservative, hesitant to reduce the primary air volume to the theoretical lower limit, thus failing to fully tap the energy-saving potential. This contradiction stems from the inability to quantitatively assess "how much the primary air volume can be reduced to at least while ensuring sufficient jet flow." Therefore, this embodiment creatively introduces and calculates in real-time the comprehensive evaluation index of "primary air ratio."

[0103] Specifically, in this embodiment, the monitoring parameter data is converted into the primary wind ratio in real time based on the principles of fluid mechanics and thermodynamics. The specific implementation method is as follows:

[0104] S231. Based on the principle of "determining air volume by coal", and referring to the theoretical air volume calculation formula for solid fuels in GB / T 26281-2021:

[0105]

[0106] Among them, the solid fuel theoretical air algorithm links the fuel quantity and the combustion air quantity through calorific value. This represents the excess air coefficient. Multiplying the theoretical air volume by the excess air coefficient yields the total air volume actually required, taking into account the risk of incomplete combustion. The leakage coefficient is calculated by subtracting the amount of air lost due to system insufficiency from the total required air volume. The remaining amount is the theoretically total combustion air volume (Nm³) including both primary and high-temperature secondary air. 3 / h).

[0107] The primary air volume represents the cost portion provided by the air cooler, which consumes electricity and reduces flame temperature. The high-temperature secondary air volume should be provided by the grate cooler, which recovers waste heat from combustion and is a free heat source.

[0108] Among them, the excess air coefficient and the air leakage coefficient are process control parameters that can be determined experimentally. They are usually set by the operator based on the composition of the kiln tail flue gas (such as oxygen content) and the sealing condition of the system, or given by the combustion control model in the DCS. This embodiment does not limit these parameters.

[0109] S232, Primary air ratio = Primary air volume / Actual total combustion air volume × 100%.

[0110] Therefore, the technical concept of this solution lies in using a theoretical model of "coal-based airflow" to back-calculate the high-temperature secondary air volume, which is difficult to measure directly, in real time. This calculation is then combined with the directly measured primary air volume to dynamically synthesize the key energy efficiency and environmental performance indicator, "primary air ratio." The core idea is to quantitatively evaluate the efficiency of primary air utilization within the entire combustion air system, shifting the operational objective from controlling isolated airflow parameters to optimizing a proportional parameter with clear physical meaning and energy efficiency implications.

[0111] Understandably, the improved approach of this solution shifts from "isolated control of primary air volume" to "coordinated optimization of the ratio between primary and secondary air." This embodiment no longer treats primary air as an independent operational variable, but rather as a component of the entire combustion air supply system for examination and control. The improvement focuses on filling the gap in real-time data on high-temperature secondary air volume through theoretical modeling and online data fusion, thereby making the higher-level optimization objective of "primary air ratio" observable and controllable.

[0112] Understandably, the "primary air ratio" directly quantifies the share of cost air (i.e., primary air volume) in the total air volume. Reducing the primary air ratio means maximizing the utilization of the enthalpy of high-temperature secondary air while meeting the total oxygen demand, reducing the heat required to heat the ambient-temperature primary air, thereby directly reducing clinker heat consumption and suppressing the formation of thermal NOx due to the potential decrease in flame peak temperature. Simultaneously, the primary air ratio is also related to jet flow rate, which needs to be maintained above the lower limit to ensure flame stability. Therefore, operation and optimization have a clear and quantifiable energy-saving approach. Operators can proactively and with evidence try to gradually reduce the primary air ratio at the edge of stable operation, exploring the lowest feasible value under current operating conditions. This helps to continuously reduce the system's coal consumption per ton of clinker and effectively suppress NOx formation, achieving a win-win situation for both economic and environmental benefits.

[0113] It is understandable that the innovation of the real-time calculation scheme for the primary air ratio in this embodiment lies not in applying a known theoretical formula, but in creatively combining online real-time data with a combustion theory model. Under the complex operating conditions of cement production, it achieves for the first time continuous, reliable, and engineered online measurement and monitoring of this core energy efficiency and environmental parameter. The primary air ratio elevates the regulation of primary air from an isolated operation based on local experience to a quantifiable and traceable refined process management activity based on global energy efficiency optimization goals. It serves as a crucial bridge connecting burner fluid dynamics regulation and overall system thermal optimization.

[0114] In existing cement firing operations, operators adjust the flame shape by regulating the valves of the swirling and axial airflows. However, the core fluid dynamics parameter determining the flame shape, "swirling intensity," cannot be directly observed or measured. Operators can only infer the appropriateness of the swirling intensity based on macroscopic phenomena such as the "thickness" or "longness" of the flame inside the kiln, combined with lagging parameters such as kiln tail temperature and kiln lining condition. This "seeing the effect and adjusting the cause" approach in existing technology suffers from severe lag and blindness. Therefore, this embodiment creatively introduces and calculates in real time the comprehensive, dimensionless evaluation index of "swirling ratio."

[0115] Specifically, in this embodiment, the monitored parameter data is converted into swirl ratio in real time based on the principles of fluid mechanics and thermodynamics. The specific implementation method is as follows:

[0116] S241. Based on the geometry of the swirl duct (such as the swirl blade angle and nozzle shape) and the measured volumetric flow rate or dynamic pressure, calculate the tangential velocity component of the airflow according to fluid mechanics principles. (Unit: m / s).

[0117] Optionally, in this embodiment, the tangential velocity component of the airflow can be obtained based on the calculation of the swirl blade angle (the most direct method) or based on the calculation of dynamic pressure measurement. This embodiment does not limit this.

[0118] For example, taking the calculation based on the swirl blade angle as an example, the specific implementation method is as follows:

[0119] a. Determine the swirl angle: The burner structural parameters include the geometric installation angle α of the swirl vanes within the swirl duct, obtained from the burner design drawings or specifications. This is a fixed structural parameter;

[0120] b. The cyclone volumetric flow rate (m³ / s) measured in real time using an online flow meter 3 / s) divided by the known cross-sectional area of ​​the vortex duct outlet (m²) 2 The average axial velocity of the airflow at the outlet of the swirl duct was calculated. ;

[0121] c. Under ideal conditions (ignoring flow losses and incomplete blade guidance), the tangential velocity component of the airflow With axial velocity component Satisfies geometric relations: . This represents the degree of "torsion" determined by the blade angle. The larger the value, the greater the axial velocity. The tangential velocity generated below The larger the volume, the stronger the vortex.

[0122] For example, taking calculations based on dynamic pressure measurement as an example, the specific implementation method is as follows:

[0123] a. Measuring tangential dynamic pressure: Install a sensor (such as a three-dimensional pitot tube) capable of measuring directional dynamic pressure in the swirl duct to obtain the dynamic pressure of the airflow in the tangential direction (unit: Pa).

[0124] b. Calculate based on the formula relating dynamic pressure and velocity: Wherein, airflow density (unit: kg / m³).

[0125] S242. Calculate the product of the rotating air mass flow rate and the tangential velocity component of the airflow to obtain the radial impulse (unit: N).

[0126] S243. Obtain the burner structural parameters, including the equivalent radius of the swirl duct. The equivalent radius of the swirl duct is a characteristic length parameter related to the burner's swirl duct structure, used to simplify and characterize the average rotational effect of the rotating airflow in theoretical calculations. In fluid mechanics and engineering practice, existing technologies have known empirical proportional relationships between the equivalent radius of the swirl duct and the actual geometric radius (such as the nozzle radius) for specific types (e.g., volute, blade) and sizes of swirlers, to reflect the non-uniformity of velocity distribution. This embodiment will not elaborate further on this.

[0127] S244. Calculate the product of the radial impulse and the equivalent radius of the swirl channel to obtain the rotational torque of the airflow in the tangential direction.

[0128] S245. Calculate the sum of the primary air momentum of each axial duct to obtain the total axial impulse. (Unit: N).

[0129] S246. Based on the rotating air mass flow rate and the total axial impulse, the equivalent diameter De of the pulse is calculated and can be expressed as:

[0130]

[0131] in, Indicates the mass flow rate of the rotating air. This indicates the airflow density.

[0132] S247. Calculate the product of the total axial impulse and the equivalent diameter of the pulse to obtain the axial torque of the airflow in the axial direction.

[0133] S248, Swirl ratio = Rotational torque of airflow in the tangential direction / Axial torque of airflow in the axial direction.

[0134] As can be seen, the swirl ratio in this embodiment quantifies the rotational intensity of the combustor outlet airflow by measuring the "disturbance" ability of the rotating airflow on the axial airflow. The larger the swirl ratio, the more dominant the rotational momentum, and the larger the size and intensity of the recirculation zone (used for entraining high-temperature flue gas for stable combustion), which can significantly enhance the mixing of fuel and combustion air. Specifically, in a high swirl ratio, the rotational torque dominates, and the airflow rotates violently and diffuses rapidly, thereby producing a short and thick flame with a large divergence angle, which is beneficial for early and rapid mixing; in a low swirl ratio, the axial torque dominates, and the airflow does not diffuse easily, thereby producing a thin and long flame with sufficient rigidity, which is beneficial for extending the fuel residence time.

[0135] Therefore, the technical concept of this solution is to link the microscopic characteristics of the flow field (i.e., swirl intensity) that cannot be directly measured with the principles of fluid mechanics, and the monitoring parameters that can be continuously measured online, as well as the fixed structural parameters of the burner (nozzle area, swirl angle, etc.). The core idea is to use readily available boundary conditions to calculate and "translate" the essential characteristics of the internal flow field.

[0136] Understandably, the improved approach of this solution represents a paradigm shift from experience-based adjustment based on outward appearances to quantitative control based on underlying causes. This embodiment no longer relies on subjective judgment of flame shape, but directly calculates and monitors the root cause (swirl ratio) that leads to the formation of that shape. This provides operators with precise numerical targets for adjusting the swirl / axial flow valves, thereby fundamentally addressing the serious lag and blindness inherent in existing adjustment techniques.

[0137] It is understandable that the innovation of the real-time calculation scheme for the swirl ratio in this embodiment lies not in the formula itself, but in combining the theoretical formula with an online sensor network and a real-time calculation system, and applying it to the specific industrial scenario of a cement rotary kiln burner. For the first time, it has achieved online quantitative perception and monitoring of the core operating variable (swirl intensity), thus providing a key data support for the intelligent and precise adjustment of the entire burner.

[0138] In order to detect slow changes in the kiln lining or abnormalities occurring at night in a timely manner, this embodiment creatively introduces and calculates the comprehensive evaluation index of "kiln lining characteristic parameters" in real time.

[0139] Specifically, in this embodiment, the monitoring parameter data is converted into kiln lining characteristic parameters in real time based on fluid mechanics and thermodynamics principles, including:

[0140] S251. Temperature calibration is performed on the infrared image of the kiln cylinder to obtain the first temperature matrix.

[0141] The method for temperature calibration of the kiln shell infrared image is existing technology. For example, the inherent chromaticity bar in the kiln shell infrared image is read; this chromaticity bar is a temperature-color mapping standard. An existing program is used to convert the chromaticity bar into RGB color data, and then clustered and compared with the color of each pixel in the kiln shell infrared image, thereby inversely mapping the pixel color to a specific temperature value. Through this step, a 1160-pixel (kiln length direction) × 380-pixel (kallature circumference direction) infrared image can be converted into a first temperature matrix of the same size, where each cell stores a temperature value (unit: °C).

[0142] S252. Average and reduce the dimensionality of the temperature values ​​in the circumferential direction of the cylinder in the first temperature matrix to obtain the second temperature matrix.

[0143] Specifically, this embodiment takes into account the large amount of data in the first temperature matrix of 1160×380 and the presence of redundant information (multiple points along the circumference of the same cross section) when directly processing it. Therefore, dimensionality reduction processing is performed first. For example, the circumferential direction of the cylinder (380 rows in the matrix) is divided into 10 equal parts, and the average temperature of all pixels in each part is taken. This operation condenses the circumferential direction information of the cylinder at each kiln length position (corresponding to the 1160th row of the matrix) into 10 feature values ​​representing the average temperature of that cross section, thus obtaining the second temperature matrix.

[0144] S253. Map the pixel coordinates of the kiln length direction in the second temperature matrix to the physical dimensions to obtain the third temperature matrix.

[0145] Specifically, the image pixel coordinates are associated with the actual physical dimensions. For example, 1160 pixels along the kiln's length are mapped to the actual length of the kiln (0-78 meters). This results in an engineered temperature matrix with 1160 rows (kiln length direction) × 10 columns (average temperature along the kiln's circumference), denoted as the third temperature matrix. In the third temperature matrix, each row represents a physical length position of the kiln, and each column represents the average temperature of a sector along the kiln's circumference at that position.

[0146] S254. The length of the kiln lining coverage area is obtained by calculating the third temperature matrix using the RSI trend analysis algorithm.

[0147] Understandably, this embodiment introduces the RSI (Relative Strength Index) trend analysis method. The entire calculation process is a process from a two-dimensional temperature field to a one-dimensional spatial sequence, and then to key point identification. First, the short-period and long-period moving averages of the temperature sequence are calculated. Then, by judging the intersection of the two moving averages, the "inflection point" where the temperature distribution changes sharply is accurately identified. This inflection point corresponds to the boundary between the firing zone and the transition zone, thereby objectively determining the endpoint of the kiln lining. Based on the determined kiln lining endpoint, the actual length of the kiln lining coverage (unit: m) is calculated, thus overcoming the subjectivity and ambiguity of human visual interpretation.

[0148] Specifically, in this embodiment, the RSI trend analysis algorithm is used to calculate the length of the kiln lining coverage area, including:

[0149] a. The temperature values ​​of the kiln-covered area along the kiln length in the third temperature matrix are denoted as a one-dimensional temperature sequence.

[0150] Exemplarily, the rows of the third temperature matrix represent the positions of the kiln length (0 - 78 meters), and the columns represent the average temperatures of 10 sectors in the circumferential direction of the cylinder. To represent the temperature level of the entire cylinder circumference with a one-dimensional temperature sequence, the average value of the temperature values of the 10 sectors corresponding to each row (i.e., each kiln length position) is calculated to form a one-dimensional temperature sequence along the kiln length. This one-dimensional temperature sequence reflects the distribution of the average temperature on the cylinder surface from the kiln head to the kiln tail.

[0151] b. To eliminate the interference of temperature random fluctuations (noise) and clearly capture the macroscopic trend of temperature changes, moving average sequences with two different periods are calculated.

[0152] On the one hand, the one-dimensional temperature sequence is divided according to a preset short period to obtain a short-period moving average sequence.

[0153] Exemplarily, taking the period N1 = 12 as an example, it represents smoothing the temperature within a kiln length range of about 0.8 meters (i.e., (12 / 1160)*78 meters). The short-period moving average sequence SMA_12[] is sensitive to temperature changes.

[0154] On the other hand, the one-dimensional temperature sequence is divided according to a preset long period to obtain a long-period moving average sequence.

[0155] Exemplarily, taking N2 = 26, it represents smoothing the temperature within a kiln length range of about 1.74 meters. The long-period moving average sequence SMA_26[] reflects a longer-term and more stable temperature trend and is insensitive to short-term fluctuations.

[0156] c. Scan the short-period moving average sequence and the long-period moving average sequence respectively in the order of temperature values from high (usually the high-temperature end at the kiln head) to low (usually the low-temperature end at the kiln tail), and perform cross-condition judgment during the scanning process to determine the end of the kiln skin.

[0157] It can be understood that this step aims to find the transition point from the high-temperature, fluctuating region (thin kiln skin) to the low-temperature, stable region (thick kiln skin).

[0158] Among them, the cross-condition is: at position j, SMA_12[j] > SMA_26[j] (that is, the short-period moving average is above the long-period moving average), and at position j + 1, SMA_12[j + 1] < SMA_26[j + 1] (the short-period moving average crosses below the long-period moving average). When this cross-condition is met, the kiln length position j is marked as a preliminary cross point, indicating that near this cross point, the short-term trend has changed from strong to weak and fallen below the long-term trend line.

[0159] Furthermore, since multiple initial crossover points may arise due to local temperature fluctuations, it is necessary to filter out false signals using validity verification rules to find the first stable trend reversal point. For example, after the initial crossover point j, it can be required that SMA_12 remains stably below SMA_26 for at least M consecutive points (e.g., M=5, approximately 0.34 meters), rather than immediately crossing back. This ensures the continuity of the trend reversal, rather than a short-lived fluctuation. Alternatively, near the crossover point, the difference between SMA_12 and SMA_26 should exceed a minimum threshold (e.g., 2°C) to avoid misjudgments caused by minor calculation errors. This embodiment does not limit this.

[0160] Starting from the kiln head, find the first pixel i_p at the kiln length position that simultaneously satisfies the cross condition and the above validity verification rules; this is the position of the end of the kiln skin.

[0161] d. Length of the kiln lining covered area = (i_p / 1160)*78.

[0162] It is understandable that this embodiment creatively applies the RSI trend analysis algorithm, which identifies trend reversals in financial time series analysis, to industrial space temperature series analysis. By utilizing the relative changes in short-term and long-term trend lines, it can keenly capture the essential change points in the temperature distribution gradient, and realize automated, high-precision, and digital measurement of the length of the kiln lining coverage area, providing key and reliable input parameters for burner adjustment and kiln condition assessment.

[0163] S255. Based on the relative temperature change of the kiln lining coverage area in the third temperature matrix, determine the kiln lining thickness characterization.

[0164] It is understandable that the kiln lining is a heat insulation layer attached to the inner wall of the kiln cylinder. For a fixed kiln internal thermal state (flame temperature and material temperature are basically stable), the temperature of the outer wall of the cylinder at a certain location is approximately negatively correlated with the thickness of the kiln lining. The thicker the kiln lining, the lower the temperature of the outer wall of the cylinder at the same location. This embodiment does not limit the specific implementation method for determining the kiln lining thickness.

[0165] For example, the method for determining the kiln lining thickness characterization in this embodiment can be as follows: The temperature value at each location within the kiln lining coverage area in the third temperature matrix is ​​directly regarded as an indirect characterization of the kiln lining thickness at that location. Next, the uniformity of distribution is determined through horizontal comparison in the spatial dimension (by comparing the temperature at different locations at the same time with the standard deviation of that location), and the trend of thickness evolution is determined through vertical comparison in the time dimension (by comparing the current temperature value with the historical baseline temperature curve or range under healthy conditions). Finally, a comprehensive and quantitative assessment of the kiln lining insulation status is formed. Kiln lining thickness characterization includes kiln lining thickening, relatively stable kiln lining thickness, and kiln lining thinning.

[0166] S256. Based on the temperature values ​​of the kiln lining coverage area in the third temperature matrix along the circumference of the cylinder and the length of the kiln, the kiln lining smoothness is calculated.

[0167] The kiln lining coverage area is comprehensively evaluated by calculating the standard deviation of the kiln circumference and the frequency of temperature fluctuations along the kiln length in the third temperature matrix. The smaller the fluctuation, the smoother the surface.

[0168] For example, in this embodiment, the kiln lining flatness is calculated based on the temperature values ​​of the kiln lining coverage area in the third temperature matrix along the circumference of the cylinder and the length of the kiln.

[0169] a. For each row of the kiln lining coverage area in the third temperature matrix, calculate the standard deviation of the 10 temperatures along the circumference of the cylinder corresponding to that row. The smaller the standard deviation, the more uniform the kiln lining thickness of that section. Calculate the mean of the standard deviations of all rows within the entire kiln lining coverage area as the matrix longitudinal flatness.

[0170] b. Locate the peaks and troughs of the temperature sequence along the kiln length in the third temperature matrix for the kiln lining covered area, and calculate the fluctuation frequency to obtain the matrix lateral flatness.

[0171] For example, a simplified method is to calculate the number of sign changes in the first-order difference. Assume a temperature sequence [215, 218, 216, 220, 218, 214, 212, 215]. Calculating the difference yields: [3, -2, 4, -2, -4, -2, 3]. Therefore, the number of sign changes (positive to negative or negative to positive) includes: (+3 to -2) 1 time, (-2 to +4) 1 time, (+4 to -2) 1 time, (-2 to -4) 0 times, (-4 to -2) 0 times, (-2 to +3) 1 time. A total of 4 changes. The more changes, the higher the fluctuation frequency and the worse the smoothness.

[0172] c. Weighted fusion of the matrix longitudinal flatness and matrix transverse flatness yields the kiln skin flatness.

[0173] For example, the reciprocal (or standardized) of the mean of the longitudinal standard deviation is weighted and fused with the reciprocal (or standardized) of the transverse fluctuation frequency to generate a kiln skin smoothness value between 0 and 1. The higher the kiln skin smoothness value, the better the smoothness.

[0174] S257, the length of the kiln skin coverage area, the kiln skin thickness, and the kiln skin flatness constitute the kiln skin characteristic parameters.

[0175] It is understandable that this embodiment outputs accurate quantitative results of kiln skin length, thickness, and flatness (index) in real time through image processing and RSI trend analysis algorithms, completely eliminating the judgment differences between different operators and establishing a unified and objective standard for the quantification of kiln skin condition. At the same time, since the length and flatness of the kiln skin are directly related to the flame shape and thermal intensity of the burner, converting these kiln skin characteristic parameters from qualitative descriptions to quantitative indicators helps to establish clear causal relationships in the future.

[0176] In summary, this embodiment creatively transforms implicit key parameters such as thermal thrust and swirl intensity, which are difficult to measure directly, into explicit indicators such as primary air momentum, thrust rate, primary air ratio, and swirl ratio, which can be calculated in real time, by processing massive amounts of monitoring parameter data in real time. This directly solves the problems of key performance parameters not being directly obtainable, not being able to be continuously monitored in real time, and not being able to dynamically adjust the operation of the kiln head burner, and transforms the thermal performance control of the kiln head burner from "post-event sampling inspection" to "online inspection".

[0177] S3. Construct an indicator analysis model. Input the core performance indicators into the indicator analysis model. The indicator analysis model outputs the burner comprehensive score obtained based on the fusion of core performance indicators and the real-time control strategy obtained based on the deviation analysis of each core performance indicator from the corresponding preset ideal threshold.

[0178] Specifically, in this embodiment, the index analysis model outputs a comprehensive burner score based on the fusion of core performance indicators, including:

[0179] S311 uses core performance indicators and smoke chamber process parameters as evaluation indicators, and sets scoring functions and corresponding initial weights for the evaluation indicators.

[0180] The scoring function set for the evaluation index can be a progressive scoring function based on fuzzy logic or a robust scoring function based on probability statistics. This embodiment does not limit this.

[0181] In this embodiment, the method for setting the initial weights for each evaluation index is not limited, as long as the sum of the initial weights for each evaluation index is 1. The Analytic Hierarchy Process (AHP) can be used, where process experts are organized to compare each evaluation index pairwise to determine its relative importance, and a scientific weight vector is derived through matrix calculation. This is the most commonly used method. Alternatively, it can be calibrated through field tests. Under stable operating conditions, a core state parameter (such as swirl ratio) can be intentionally changed, and its sensitivity to the final target (such as clinker strength and coal consumption) can be observed. Initial weights can then be allocated based on the magnitude of the sensitivity.

[0182] For example, primary air momentum directly determines the most fundamental parameter of flame penetration and rigidity, and is the basis for adjustment; therefore, its initial weight is set to 0.18. Thrust rate, as an energy efficiency supplement to primary air momentum, evaluates the efficiency of momentum generation; therefore, its initial weight is set to 0.07, lower than momentum, reflecting its supplementary and optimization attributes. Swirl ratio controls flame shape, mixing intensity, and thermal distribution, and has a direct and rapid impact on kiln lining formation; it is a key adjustment variable, therefore, its initial weight is set to 0.2. Primary air ratio is directly related to coal and electricity consumption, and is a key optimization direction for energy saving and consumption reduction; therefore, its initial weight is set to 0.09. Kiln lining length is a direct and macroscopic manifestation of the burner's adjustment effect within the kiln; abnormal length directly affects output and quality stability; therefore, its initial weight is set to... The initial weight is set at 0.15; the kiln lining thickness directly reflects the protective state of the kiln lining, and thinning of the lining is a direct precursor to "red kiln" accidents, requiring continuous monitoring, so the initial weight is set at 0.08; the kiln lining flatness reflects the uniformity and stability of combustion, and poor flatness indicates local overheating or flame erosion, so the initial weight is set at 0.07; the kiln tail flue temperature reflects the overall calcination temperature level and is a comprehensive result parameter of thermal regime stability, so the initial weight is set at 0.06; the flue gas oxygen concentration reflects the excess air coefficient, affecting thermal efficiency and NOx generation, so the initial weight is set at 0.05; the flue gas carbon monoxide concentration reflects the degree of incomplete combustion, and excessively high concentrations pose safety risks and heat loss, so the initial weight is set at 0.05.

[0183] Furthermore, before S312, outlier removal, missing value filling, smoothing filtering, and standardization operations can be performed on each evaluation index to obtain preprocessed evaluation indexes.

[0184] S312: Calculate the current values ​​of each evaluation indicator based on the scoring function, and then perform standardization to obtain the current score value of each evaluation indicator.

[0185] Optionally, a progressive scoring function based on fuzzy logic is used to calculate the current values ​​of each evaluation indicator, followed by standardization to obtain the current score value for each indicator. Specifically, this is achieved by abandoning hard thresholds and using membership functions from fuzzy logic to define the score. For each evaluation indicator, fuzzy sets such as "Excellent," "Good," "Average," and "Poor" are defined, and smooth membership functions (such as triangular, trapezoidal, or Gaussian functions) are designed for them. The score value is the membership degree of the current value of the evaluation indicator relative to the "Excellent" set, mapped to a score of 0-100.

[0186] For example, let's take the calculation of the thrust ratio score as an example:

[0187] Define the interval [8,11]N / MW as "excellent" and use a trapezoidal membership function, where the membership degree is 1 (the score is mapped to 100 points) within this interval.

[0188] The intervals [7,8) and (11,12] are defined as "good", and the membership degree decreases linearly from 1 to 0.8 (the score decreases linearly from 100 to 80).

[0189] The intervals [6,7) and (12,13] are defined as "medium". A sloped membership function is used, with the membership degree linearly decreasing from 0.8 to 0.6 (the score value linearly decreasing from 80 to 60).

[0190] A score below 6 N / MW or above 13 N / MW is defined as "poor", and the membership level drops rapidly to 0 (the score is mapped to 0).

[0191] The incremental scoring function based on fuzzy logic calculates the current values ​​of each evaluation indicator, which makes the score changes continuous and smooth, more in line with process cognition, and avoids score jumps caused by hard thresholds.

[0192] Optionally, a robustness scoring function based on probability statistics is used to calculate the current values ​​of each evaluation indicator, followed by standardization to obtain the current score value for each indicator. Specifically, the score is not based on the deviation of the current value of the evaluation indicator from a fixed threshold, but rather on the probability of the current value's position within the historical best operating condition data distribution. Assuming that a certain evaluation indicator follows a certain probability distribution (such as a normal distribution) under the historical best operating conditions, the score of the current indicator is the cumulative probability value (or likelihood value) under that probability distribution, mapped to a score between 0 and 100.

[0193] For example, consider the calculation of the score for smoke chamber oxygen concentration: Collect all smoke chamber oxygen concentration data from the past month during periods when both clinker quality and energy consumption met standards, and fit its historical best probability distribution. The score for the current smoke chamber oxygen concentration is equal to the percentile (or transformed likelihood value) of the current smoke chamber oxygen concentration in the historical best probability distribution. For example, if the current smoke chamber oxygen concentration is at the median of the historical best probability distribution, the score is approximately 90 points (considering the central tendency of the distribution); if it is at the extreme tail of the distribution, the score is approximately 0 points.

[0194] The robustness scoring function based on probability statistics has solid statistical significance in calculating the current values ​​of each evaluation index. It can reflect the statistical characteristics of whether the current operating condition is "like" the best period in history, thus making it more robust to noise and occasional fluctuations.

[0195] S313, the current scores of each evaluation indicator and their corresponding initial weights are weighted and summed to obtain the comprehensive score of the burner.

[0196] Optionally, in this embodiment, the real-time control strategy can be implemented based on the analysis of the deviation between each core performance indicator and its corresponding preset ideal threshold, or it can be implemented based on the analysis of the deviation between the current score value of the evaluation indicator and its corresponding preset ideal score value using preset fixed rules. An example is shown below:

[0197] 1. If the primary air momentum score is less than 60 and the thrust rate score is less than 60, a real-time control strategy will be pushed: "The primary air momentum and thrust rate are too low, and the flame rigidity is insufficient. Recommendation: Check and increase the primary air volume, or slightly reduce the amount of coal at the kiln head."

[0198] 2. If the swirl ratio score is <60 and the kiln tail smoke chamber temperature score is <60, a real-time control strategy will be pushed: "The swirl ratio is too high and the firing zone temperature exceeds the limit, which may indicate flame scouring. Recommendation: Appropriately reduce the opening of the swirl damper and optimize the swirl intensity."

[0199] 3. If the kiln lining length score is less than 60, a real-time control strategy will be pushed: "Kiln lining length abnormal (too short / too long). Recommendation: Adjust the ratio of internal and external air in the burner and the position of the coal pipe in accordance with the prompt to optimize the flame length."

[0200] Furthermore, in this embodiment, the acquired monitoring parameter data, the converted core performance indicators, and the fused burner comprehensive score are all displayed in numerical or parameter curve form on the visualization panel. The analyzed real-time control strategies are also displayed in a real-time list of specific and actionable text suggestions on the visualization panel for direct operator reference and execution. (Appendix) Figure 2 This is a first visual panel diagram illustrating the thermal performance of the kiln head burner according to an embodiment of the present invention. (Attached) Figure 3 This is a schematic diagram of the second visualization panel for the thermal performance of the kiln head burner in an embodiment of the present invention.

[0201] Understandably, because the index analysis model encodes implicit and intuitive "experience" into explicit and digital "rules," it establishes a unified and objective quantitative standard for evaluating the thermal performance of burners. This provides operators with a reliable data foundation for controlling the kiln head burners, enabling the standardization, digital accumulation, and lossless inheritance of excellent process knowledge.

[0202] S4 controls the kiln head burner based on a real-time control strategy and continuously collects the core performance indicators after control.

[0203] The real-time control strategy output in step S3 can be manually executed by the operator. After the control action is executed, the system initiates a preset monitoring cycle. Within this cycle, the system synchronously and in time alignment collects the core performance indicators after the control. This preset monitoring cycle takes into account the response lag characteristics of the kiln system.

[0204] S5 performs causal correlation analysis on the core performance indicators after regulation, and iteratively optimizes the fusion strategy of the indicator analysis model based on the causal correlation analysis results.

[0205] Specifically, in this embodiment, causal correlation analysis is performed on the core performance indicators after regulation, and the fusion strategy of the indicator analysis model is iteratively optimized based on the causal correlation analysis results, including:

[0206] S51 uses the adjusted primary air momentum, thrust rate, primary air ratio, and swirl ratio as dependent variables, and the adjusted kiln skin characteristic parameters and smoke chamber process parameters as resultant variables.

[0207] S52 performs time alignment on the dependent and result variables within the preset monitoring period, and combines the real-time control strategy, control timestamp, dependent variable, and result variable to form a data sample.

[0208] Understandably, weight optimization is not performed in real time, but rather at fixed, relatively long monitoring intervals (e.g., every 24 hours, every week, or after accumulating 500 new samples). This ensures that there is sufficient data for subsequent training of the machine learning model and calculation of influence coefficients, resulting in stable conclusions.

[0209] S53, extract features from the data samples to obtain input features and target variables.

[0210] The data samples were preprocessed to extract key features. For example, "reduction in swirl ratio" and "rate of change in primary air momentum" were used as input features, while "change in kiln lining smoothness" and "variance of average smoke chamber temperature" were used as target variables.

[0211] S54. Construct a machine learning model. Train the machine learning model based on the input features and the target variable. The output of the trained machine learning model is the influence coefficient of each dependent variable on the result variable.

[0212] This can be achieved by training machine learning models such as random forest regression or linear regression, using input features as input, to predict the target variable after regulation. The method of training the machine learning model based on input features and the target variable in this embodiment is similar to the existing method of training the model based on sample data, and is a conventional technique used by those skilled in the art; therefore, it will not be elaborated further in this embodiment.

[0213] Among them, the feature importance scores (random forest regression model) or regression coefficients (linear regression model) of the trained machine learning model intuitively reveal the direction (positive or negative sign) and intensity of the quantitative influence of each dependent variable on the result variable, i.e., the "influence coefficient". For example, the trained machine learning model may show that the "swirl ratio" has the largest negative influence coefficient on "kiln skin smoothness", while the "primary air momentum" has a positive influence coefficient on "kiln skin length".

[0214] S55, based on the influence coefficients of each dependent variable on the effect variable, optimizes the initial weights corresponding to each evaluation index, and uses the optimized weights in the next burner comprehensive score fusion process.

[0215] Specifically, the implementation of S55 is as follows:

[0216] First, the influence coefficients of each dependent variable on the effect variable are normalized into a probability distribution, the sum of which is 1, thus obtaining a target weight W_target that is entirely based on the influence of data within the preset monitoring period.

[0217] Secondly, to avoid drastic weight changes due to fluctuations in single batches of data within the preset monitoring period, a smooth update strategy is adopted. The optimized weight W_new is the weighted average of the current initial weight W_current and the target weight, which can be expressed as: W_new = β * W_target + (1-β) * W_current. Here, β is the learning rate (e.g., 0.1–0.3), a hyperparameter preset according to actual conditions. A smaller β (e.g., 0.1) means that the weight changes very slowly and robustly, relying on long-term data trends; a larger β (e.g., 0.3) allows the weight to respond more quickly to recent changes in process relationships.

[0218] Finally, after obtaining W_new, the new weights of each evaluation index need to be re-standardized to ensure that the sum of each component is 1. At the same time, for the indicators that are recognized as extremely critical (such as "kiln skin length" and "primary air momentum"), their weights can be set to not be lower than the preset protection value.

[0219] Furthermore, this embodiment also allows for knowledge accumulation: monitoring parameter data, data samples of "real-time control strategy - control timestamp - dependent variable - effect variable", the influence coefficients of each dependent variable on the effect variable, and the new weights of the optimized evaluation indicators are all stored in the burner operation knowledge base. This knowledge base exists in a queryable, structured form, recording "under what operating conditions (monitoring parameter data), what adjustments (real-time control strategy, dependent variable) are taken, and what effects (effect variable) are most likely to be produced." As the operating time increases, the knowledge base becomes richer, which helps to directly output causal correlation analysis and push specific burner control strategies based on the correlation patterns in the knowledge base.

[0220] It is understood that this embodiment aims to provide operators with a set of auxiliary tools with accurate judgment criteria and reasonable adjustment paths when facing kiln surface abnormalities. In the case where operators lack a clear judgment direction, it achieves a closed loop of "analysis-suggestion-control-optimization" through data-driven approach.

[0221] Example 2:

[0222] Please see the appendix Figure 4 , Figure 4 This is a schematic diagram of the closed-loop control system for the thermal performance of a kiln head burner, provided as an embodiment of this specification.

[0223] like Figure 4 As shown, the closed-loop control system for the thermal performance of the kiln head burner may include at least:

[0224] Data monitoring module 1 is used to acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared image and fuel parameters;

[0225] The index calculation module 2 is used to convert monitoring parameter data into core performance indicators in real time based on the principles of fluid mechanics and thermodynamics. The core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters.

[0226] The comprehensive analysis module 3 is used to construct an indicator analysis model. The core performance indicators are input into the indicator analysis model, and the indicator analysis model outputs a comprehensive score of the burner based on the fusion of the core performance indicators, as well as a real-time control strategy based on the analysis of the deviation of each core performance indicator from the corresponding preset ideal threshold.

[0227] The control module 4 is used to control the kiln head burner based on a real-time control strategy and continuously collect the core performance indicators after control.

[0228] Feedback optimization module 5 is used to perform causal correlation analysis on the core performance indicators after regulation, and to iteratively optimize the fusion strategy of the indicator analysis model based on the causal correlation analysis results.

[0229] Furthermore, in this embodiment, the closed-loop control system for the thermal performance of the kiln head burner also includes:

[0230] The visualization module is used to visualize monitoring parameter data, core performance indicators, burner comprehensive score, and real-time control strategies.

[0231] It is understood that the technical concept of the closed-loop control system for the thermal performance of the kiln head burner provided in this embodiment is similar to the technical concept of the closed-loop control method for the thermal performance of the kiln head burner mentioned above, and will not be repeated here.

[0232] The above description is merely a preferred embodiment disclosed in this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of protection involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this disclosure.

[0233] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

Claims

1. A closed-loop control method for the thermal performance of a kiln head burner, characterized in that, Includes the following steps: Acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared images, and fuel parameters; Based on the principles of fluid mechanics and thermodynamics, the monitoring parameter data is converted into core performance indicators in real time. These core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters. A performance index analysis model is constructed. Core performance indexes are input into the performance index analysis model. The performance index analysis model outputs a comprehensive burner score based on the fusion of core performance indexes and a real-time control strategy based on the analysis of the deviation of each core performance index from the corresponding preset ideal threshold. The kiln head burner is controlled based on a real-time control strategy, and the core performance indicators after control are continuously collected. Causal correlation analysis was conducted on the core performance indicators after regulation, and the fusion strategy of the indicator analysis model was iteratively optimized based on the results of the causal correlation analysis.

2. The closed-loop control method for the thermal performance of the kiln head burner as described in claim 1, characterized in that, The fuel parameters include the kiln head coal consumption and the lower heating value of pulverized coal. Based on fluid mechanics and thermodynamics principles, the monitoring parameter data is converted into core performance indicators in real time, including: The kiln heat load is obtained by multiplying the kiln head coal consumption by the lower heating value of pulverized coal and dividing by the unit clinker output. The ratio of primary air velocity to kiln heat load is calculated to obtain the thrust rate, which is used to measure the energy utilization efficiency of the burner.

3. The closed-loop control method for the thermal performance of the kiln head burner as described in claim 1, characterized in that, The airflow parameters include the rotary air mass flow rate. Based on fluid mechanics and thermodynamics principles, the monitored parameter data is converted into core performance indicators in real time, including: The tangential velocity component of the airflow is calculated based on the principles of fluid mechanics. The radial impulse is obtained by calculating the product of the rotating air mass flow rate and the tangential velocity component of the airflow. Obtain the burner structural parameters, including the equivalent radius of the swirl duct; The product of radial impulse and the equivalent radius of the swirl channel is calculated to obtain the rotational torque of the airflow in the tangential direction; Calculate the sum of the primary wind momentum of each axial duct to obtain the total axial impulse; The equivalent diameter of the pulse is calculated based on the mass flow rate of the rotating air and the total axial impulse. The axial moment of the airflow in the axial direction is obtained by calculating the product of the total axial impulse and the equivalent diameter of the pulse. The ratio of the rotational torque of the airflow in the tangential direction to the axial torque in the axial direction is calculated to obtain the swirl ratio, which is used to quantify the rotational intensity of the airflow at the burner outlet.

4. The closed-loop control method for the thermal performance of the kiln head burner as described in claim 1, characterized in that, The air volume parameter includes primary air volume. Based on fluid mechanics and thermodynamics principles, the monitored parameter data is converted into core performance indicators in real time, including: Based on fuel parameters, the actual total combustion air volume, including primary air volume and high-temperature secondary air volume, is calculated using the solid fuel theoretical air algorithm. The ratio of primary air volume to actual total combustion air volume is calculated and converted into a percentage to obtain the primary air proportion, which is used to quantify the balance between the dependence of the combustion process on primary air volume and the utilization efficiency of high-temperature secondary air.

5. The closed-loop control method for the thermal performance of a kiln head burner as described in claim 1, characterized in that, Based on the principles of fluid mechanics and thermodynamics, monitoring parameter data is transformed into core performance indicators in real time, including: Temperature calibration was performed on the infrared image of the kiln shell to obtain the first temperature matrix; The temperature values ​​along the circumference of the cylinder in the first temperature matrix are averaged and reduced in dimension to obtain the second temperature matrix. The pixel coordinates along the kiln length in the second temperature matrix are mapped to physical dimensions to obtain the third temperature matrix; The length of the kiln lining coverage area was obtained by calculating the third temperature matrix using the RSI trend analysis algorithm. The kiln lining thickness characterization is determined based on the relative temperature change of the kiln lining coverage area in the third temperature matrix. The kiln skin flatness is calculated based on the temperature values ​​of the kiln skin coverage area in the third temperature matrix along the circumference of the cylinder and the length of the kiln. The length of the kiln lining coverage area, the thickness of the kiln lining, and the flatness of the kiln lining constitute the kiln lining characteristic parameters.

6. The closed-loop control method for the thermal performance of a kiln head burner as described in claim 5, characterized in that, The RSI trend analysis algorithm is used to calculate the length of the kiln lining coverage area by calculating the third temperature matrix, including: The temperature values ​​of the kiln-length direction in the kiln-covered area in the third temperature matrix are denoted as a one-dimensional temperature sequence. The one-dimensional temperature sequence is divided according to a preset short period to obtain a short-period moving average sequence. The one-dimensional temperature sequence is divided according to a preset long period to obtain a long-period moving average sequence. According to the temperature values ​​from high to low, the short-period moving average sequence and the long-period moving average sequence are scanned respectively, and the cross condition judgment is performed during the scanning process to determine the end of the kiln skin. The length of the area covered by the kiln lining is calculated based on the coordinate positioning of the end of the kiln lining.

7. The closed-loop control method for the thermal performance of a kiln head burner as described in claim 5, characterized in that, Based on the temperature values ​​of the kiln lining covered area along the circumference of the kiln cylinder and along the length of the kiln in the third temperature matrix, the kiln lining smoothness is calculated, including: The mean of the temperature standard deviation of the kiln lining covered area in the circumferential direction of the cylinder in the third temperature matrix is ​​calculated to obtain the longitudinal flatness of the matrix. Calculate the temperature fluctuation frequency of the kiln lining covered area along the kiln length in the third temperature matrix to obtain the matrix lateral flatness. The kiln skin flatness is obtained by weighted fusion of the matrix longitudinal flatness and matrix transverse flatness.

8. The closed-loop control method for the thermal performance of a kiln head burner as described in claim 1, characterized in that, The monitoring parameter data also includes smoke chamber process parameters. The index analysis model outputs a comprehensive burner score based on the fusion of core performance indicators, including: The core performance indicators and smoke chamber process parameters are used as evaluation indicators, and scoring functions and corresponding initial weights are set for the evaluation indicators. The current values ​​of each evaluation indicator are calculated based on the scoring function, and then standardized to obtain the current score value of each evaluation indicator. The current scores of each evaluation indicator and their corresponding initial weights are weighted and summed to obtain the overall score of the burner.

9. The closed-loop control method for the thermal performance of a kiln head burner as described in claim 8, characterized in that, Causal correlation analysis was performed on the core performance indicators after regulation, and the fusion strategy of the indicator analysis model was iteratively optimized based on the results of the causal correlation analysis, including: The primary air momentum, thrust rate, primary air ratio, and swirl ratio after adjustment were used as dependent variables, and the kiln skin characteristic parameters and smoke chamber process parameters after adjustment were used as resultant variables. Time alignment is performed on the dependent and result variables within the preset monitoring period, and the real-time control strategy, control timestamp, dependent and result variables are combined to form a data sample; Feature extraction is performed on the data samples to obtain the input features and target variables; Construct a machine learning model, train the machine learning model based on input features and target variables, and output the machine learning model after training as the influence coefficients of each dependent variable on the outcome variable; Based on the influence coefficients of each dependent variable on the result variable, the initial weights corresponding to each evaluation index are optimized, and the optimized weights are used in the next fusion process of burner comprehensive scoring.

10. A closed-loop control system for the thermal performance of a kiln head burner, characterized in that, include: The data monitoring module is used to acquire monitoring parameter data of the kiln head burner, including air volume parameters, kiln shell infrared images, and fuel parameters. The index calculation module is used to convert monitoring parameter data into core performance indicators in real time based on the principles of fluid mechanics and thermodynamics. The core performance indicators include primary air momentum, thrust rate, primary air ratio, swirl ratio, and kiln skin characteristic parameters. The comprehensive analysis module is used to build an indicator analysis model. The core performance indicators are input into the indicator analysis model, and the indicator analysis model outputs a comprehensive score of the burner based on the fusion of the core performance indicators, as well as a real-time control strategy based on the analysis of the deviation of each core performance indicator from the corresponding preset ideal threshold. The control module is used to control the kiln head burner based on a real-time control strategy and continuously collect the core performance indicators after control. The feedback optimization module is used to perform causal correlation analysis on the core performance indicators after regulation, and to iteratively optimize the fusion strategy of the indicator analysis model based on the causal correlation analysis results.

Images