Flue gas heating advance regulation method and system based on airspeed-conversion rate driven lightgbm-lstm and storage medium
Patent Information
- Application Number
- CN202610974542.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-01
- Publication Date
- 2026-10-16
AI Technical Summary
[0006]本发明目的在于提供一种基于空速-转化率驱动LightGBM-LSTM的烟气加热超前调节方法，旨在解决烧结烟气一氧化碳催化氧化系统中因空速变化导致热惯性滞后、烟气温度波动大、传统反馈调节无法提前匹配热量需求的问题，具体技术方案如下：
本发明通过空速-转化率物理机理的嵌入，提前量化一氧化碳自放热对烟气温度的贡献，预判煤气补热需求，使控制动作能够提前于温度变化发生，从根源上克服了热惯性对控制品质的影响，有效抑制温度波动；在此基础上，通过LightGBM完成多工况下煤气需求量的高精度稳态拟合，再通过LSTM基于历史误差序列输出时序残差补偿值，专门补偿热惯性滞后、工况突变以及挥发性有机化合物放热等因素带来的预测偏差。通过将挥发性有机化合物放热引起的微小温度偏差纳入LSTM的残差补偿范围，本发明显著降低了对昂贵挥发性有机化合物在线检测设备的依赖，在保证控制精度的同时降低了系统成本，实现物理机理与数据驱动的深度融合，将控制逻辑从传统的滞后响应转变为超前预判与精准匹配。
Smart Images
Figure CN122923451A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of sintering flue gas heating control technology, specifically to a flue gas heating advance regulation method, system, and storage medium based on space velocity-conversion rate driven LightGBM-LSTM. Background Technology
[0002] The flue gas generated during the sintering process contains a large amount of carbon monoxide, which needs to be catalytically oxidized to meet emission standards. The carbon monoxide catalytic reaction is highly sensitive to temperature windows; typically, the flue gas needs to be heated and stabilized at around 280°C to ensure catalyst activity, inhibit impurity poisoning, and maintain a high conversion rate. The initial flue gas temperature is too low, requiring the combustion of blast furnace gas for supplemental heating; simultaneously, the catalytic oxidation of carbon monoxide releases heat, directly increasing the flue gas temperature and affecting the system's thermal balance.
[0003] In actual operation, frequent fluctuations in flue gas volume directly lead to significant changes in space velocity: the larger the flue gas volume, the higher the space velocity, the shorter the contact time between carbon monoxide and the catalyst, the lower the conversion rate, the less heat released in the reaction, the smaller the temperature rise of the flue gas itself, and the more gas supplementary heating is required; conversely, a smaller flue gas volume results in a lower space velocity, a higher conversion rate, a larger heat release, a significantly reduced need for supplementary heating, and may even allow for complete shutdown of gas heating. Therefore, space velocity is the core bridge connecting flue gas fluctuations and changes in heat demand.
[0004] To maintain a stable reaction temperature, existing systems typically employ a feedback control strategy based on measured temperature, adjusting the gas supply according to the deviation between the catalyst bed inlet temperature and the target temperature. However, due to the system's thermal inertia (a lag between heat input and temperature response), this lagging adjustment method cannot anticipate heat demand, often resulting in excessive carbon monoxide emissions due to low temperatures or wasted gas due to high temperatures. Building on this, industrial sites commonly employ a scheme of heating flue gas with blast furnace gas and combining it with a gas-fired heating element (GGH) to increase the inlet temperature to enhance heating capacity. However, this scheme remains a passive response adjustment and fails to address the overshoot or undershoot problems caused by flue gas fluctuations at their root.
[0005] In summary, existing technologies in sintering flue gas carbon monoxide catalytic oxidation systems have the following shortcomings: traditional feedback control cannot respond to changes in heat demand in advance due to thermal inertia lag; existing gas heating schemes are still passive adjustments and cannot match flue gas fluctuations. Therefore, there is an urgent need for a flue gas heating control method that deeply integrates the space velocity-conversion physical mechanism with intelligent prediction models to achieve proactive adjustment. Summary of the Invention
[0006] The purpose of this invention is to provide a method for advance regulation of flue gas heating based on space velocity-conversion driven LightGBM-LSTM, aiming to solve the problems of thermal inertia lag caused by space velocity changes, large flue gas temperature fluctuations, and the inability of traditional feedback regulation to match heat demand in advance in sintering flue gas carbon monoxide catalytic oxidation systems. The specific technical solution is as follows: A method for advance regulation of flue gas heating based on space velocity-conversion rate driven LightGBM-LSTM includes: The operating parameters of the sintering flue gas heating system are collected, including the flue gas volume. Inlet carbon monoxide concentration Initial flue gas temperature and catalytic reactor inlet temperature ; The collected operating parameters are preprocessed and adjusted according to the flue gas volume. With catalyst volume Calculate airspeed ; According to airspeed and inlet carbon monoxide concentration Determine whether the current operating condition is within a stable control range; If the system is within a stable control range, then the mechanism-driven anti-adjustment process will be executed: Based on airspeed Calculate the current carbon monoxide conversion rate And based on the inlet carbon monoxide concentration and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Based on the initial flue gas temperature Target temperature at the inlet of the catalytic reactor Temperature rise in smoke Calculate the gas supplementary heating demand ; like If the value is greater than 0, the current gas reference quantity is output through the LightGBM model, and the current time-series residual compensation value is output through the LSTM model. The gas reference quantity and the time-series residual compensation value are then superimposed to obtain the final gas quantity at the current moment. If If the value is ≤ 0, then the gas heating will be turned off; If the temperature is not within the stable control range, then the adaptive adjustment process for the catalytic reactor inlet temperature will be executed: Based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor The deviation is used to calculate the gas adjustment amount, and the final gas amount at the current moment is updated to be the sum of the final gas amount at the previous moment and the gas adjustment amount.
[0007] Preferably, based on the amount of flue gas With catalyst volume Calculate airspeed Represented as: Among them, catalyst volume The catalyst loading volume in the catalytic reactor; space velocity. This represents the gas volume flow rate passing through a unit volume of catalyst per unit time.
[0008] Preferably, if the inlet carbon monoxide concentration under the current operating conditions is... In the range of 3 to 10 Within range and airspeed Between 9000 and 40000 If the current condition is within the specified range, it is considered to be in a stable control range; otherwise, it is considered not to be in a stable control range.
[0009] Preferably, based on the inlet carbon monoxide concentration and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Represented as: in, The calorific value of carbon monoxide; The density of the flue gas; This refers to the specific heat capacity of the flue gas.
[0010] Preferably, the LightGBM model uses the current flue gas volume. Inlet carbon monoxide concentration Initial flue gas temperature ,airspeed and gas heating demand Given the input features, output the current gas reference quantity.
[0011] Preferably, the LSTM model takes the historical gas prediction error sequence of the LightGBM model as input and outputs the time-series residual compensation value at the current moment; Among them, the historical predicted gas error sequence consists of the predicted gas error at several past sampling times. The predicted gas error at each sampling time is equal to the difference between the actual optimal gas quantity at that time and the gas reference quantity predicted by the LightGBM model. This difference includes the temperature rise error caused by the oxidation of volatile organic compounds.
[0012] Preferably, the gas adjustment amount in the adaptive temperature control process of the catalytic reactor inlet is expressed as: in, The density of the flue gas; The specific heat capacity of the flue gas; The calorific value of the gas; For gas regulation, if If positive, then increase the final gas volume based on the previous moment. The amount of gas, if If the value is negative, then the final gas volume will be reduced from the previous moment's value. The amount of gas.
[0013] Preferably, the gas supplementary heating demand Equal to the target temperature at the inlet of the catalytic reactor Subtract the initial flue gas temperature Temperature rise in smoke .
[0014] The present invention also provides a flue gas heating advance regulation system based on space velocity-conversion rate driven LightGBM-LSTM, including a memory and a processor. The memory stores a computer program, and the processor executes the flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM when running the computer program.
[0015] The present invention also provides a storage medium storing a computer program, which, when executed by a processor, performs the aforementioned flue gas heating advance regulation method based on airspeed-conversion rate driven LightGBM-LSTM.
[0016] The application of the technical solution of the present invention has the following beneficial effects: This invention embeds the space velocity-conversion rate physical mechanism to quantify the contribution of carbon monoxide self-exothermic reaction to flue gas temperature in advance, predicting the gas reheat demand and enabling control actions to occur before temperature changes. This fundamentally overcomes the impact of thermal inertia on control quality and effectively suppresses temperature fluctuations. Based on this, LightGBM is used to achieve high-precision steady-state fitting of gas demand under multiple operating conditions. Then, LSTM outputs time-series residual compensation values based on historical error sequences to specifically compensate for prediction deviations caused by thermal inertia lag, sudden changes in operating conditions, and volatile organic compound (VOC) exothermic reactions. By incorporating the small temperature deviations caused by VOC exothermic reactions into the residual compensation range of LSTM, this invention significantly reduces reliance on expensive online VOC detection equipment, lowering system costs while maintaining control accuracy. It achieves deep integration of physical mechanisms and data-driven approaches, transforming control logic from traditional delayed response to proactive prediction and precise matching.
[0017] This invention employs a dual-mode adaptive switching strategy. Within the stable control range, it operates in a mechanism-driven precise prediction mode for fine-tuning. When parameters exceed the stable range, it automatically switches to an inlet temperature adaptive adjustment mode, ensuring stable operation across all conditions. The two modes seamlessly connect, guaranteeing both high-precision control under stable conditions and safe operation under extreme conditions, truly achieving robust control covering all operating conditions.
[0018] This invention quantifies the contribution of carbon monoxide self-exothermic heating to flue gas temperature through mechanistic calculations. When the self-exothermic heating is sufficient to maintain the target temperature, the gas heating is directly shut off, relying entirely on the heat released from the catalytic oxidation of carbon monoxide to maintain the reaction temperature, achieving zero additional energy consumption. Since the carbon monoxide concentration in sintering flue gas is usually high, the self-exothermic heating can cover most or even all of the heating requirements under most operating conditions. Therefore, this invention significantly reduces blast furnace gas consumption, achieving energy conservation and emission reduction while ensuring carbon monoxide emissions meet standards, thus possessing both environmental and economic benefits.
[0019] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description
[0020] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of the flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM in Example 1; Figure 2 This is a schematic diagram of the conveying path of sintering flue gas in Example 1. Detailed Implementation
[0021] To facilitate understanding of the present invention, a more complete description is provided below, along with preferred embodiments. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the present invention.
[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
[0023] Example 1: See Figure 1This embodiment provides a method for advance regulation of flue gas heating based on space velocity-conversion rate driven LightGBM-LSTM, applicable to flue gas temperature control in sintering flue gas carbon monoxide catalytic oxidation systems. This method integrates space velocity-conversion rate mechanism calculation with LightGBM benchmark prediction and LSTM time-series residual compensation to achieve advance regulation and precise control of flue gas heating. Specifically, it includes the following steps: The operating parameters of the sintering flue gas heating system are collected, including the flue gas volume. Inlet carbon monoxide concentration Initial flue gas temperature and catalytic reactor inlet temperature ; The collected operating parameters are preprocessed and adjusted according to the flue gas volume. With catalyst volume Calculate airspeed ; According to airspeed and inlet carbon monoxide concentration Determine whether the current operating condition is within a stable control range; If the system is within a stable control range, then the mechanism-driven anti-adjustment process will be executed: Based on airspeed Calculate the current carbon monoxide conversion rate And based on the inlet carbon monoxide concentration and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Based on the initial flue gas temperature Target temperature at the inlet of the catalytic reactor Temperature rise in smoke Calculate the gas supplementary heating demand Among them, the target temperature at the inlet of the catalytic reactor is... The optimal temperature for the catalytic oxidation of carbon monoxide is set as the inlet temperature of the catalytic reactor in this embodiment. Set to 280℃.
[0024] like If the value is greater than 0, the current gas reference quantity is output through the LightGBM model, and the current time-series residual compensation value is output through the LSTM model. The gas reference quantity and the time-series residual compensation value are then superimposed to obtain the final gas quantity at the current moment. If If the value is ≤ 0, then the gas heating will be turned off; If the temperature is not within the stable control range, then the adaptive adjustment process for the catalytic reactor inlet temperature will be executed: Based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor The deviation is used to calculate the gas adjustment amount, and the final gas amount at the current moment is updated to be the sum of the final gas amount at the previous moment and the gas adjustment amount.
[0025] The flue gas heating advance adjustment method of this embodiment will be described in detail below: like Figure 2 As shown, this embodiment collects the following operating parameters in real time using online sensors: flue gas volume That is, the flue gas flow rate after passing through the heat exchanger (GGH) in the reactor, in units of... Inlet carbon monoxide concentration This refers to the carbon monoxide concentration after passing through the heat exchanger (GGH) in the reactor, expressed in units of... Initial flue gas temperature The initial flue gas temperature after passing through the heat exchanger in the reactor is expressed in units of... ; Catalytic reactor inlet temperature This refers to the flue gas temperature above the catalyst bed, expressed in units of... Furthermore, the collected raw data undergoes preprocessing, including filtering and denoising, outlier removal, missing value imputation, and time-series synchronization processing of high- and low-frequency sampling data to eliminate signal interference and sampling asynchrony issues in industrial settings. Specifically, the time-series synchronization processing of high- and low-frequency sampling data involves aligning sensor data from different sampling frequencies to a unified time reference through interpolation, resampling, or time registration methods to ensure that all parameters used in subsequent calculations originate from the same point in time.
[0026] Preferably, based on the amount of flue gas With catalyst volume Calculate airspeed Represented as: Among them, catalyst volume The volume of catalyst loaded in the catalytic reactor is expressed in units of... ;airspeed The unit is , which represents the gas volume flow rate passing through a unit volume of catalyst per unit time, reflecting the contact time between the flue gas and the catalyst.
[0027] Preferably, if the inlet carbon monoxide concentration under the current operating conditions is... In the range of 3 to 10 Within range and airspeed Between 9000 and 40000 If the current condition is within the specified range, it is considered to be in a stable control range; otherwise, it is considered not to be in a stable control range (i.e., an unstable control range). If the current condition is in a stable control range, the mechanism-driven advance adjustment process is executed; if the current condition is not in a stable control range, the adaptive adjustment process for the catalytic reactor inlet temperature is executed.
[0028] Within the stable control range, since the carbon monoxide conversion rate decreases with increasing space velocity and increases with decreasing space velocity, the space velocity can be empirically controlled. With carbon monoxide conversion rate The relationship is represented as: in, It is a constant. Less than 0, carbon monoxide conversion rate The value ranges from 0 to 1, representing the proportion of carbon monoxide participating in the reaction to the total amount of carbon monoxide entering the inlet.
[0029] Preferably, the constant The calibration was performed as follows: Within the stable control range, carbon monoxide conversion was measured at multiple airspeed points. The measured data were then fitted using a quadratic polynomial fit employing the least squares method to obtain the desired result. The calibration cycle is determined based on changes in catalyst activity, for example, recalibrating once per quarter.
[0030] Furthermore, since the exothermic oxidation of carbon monoxide and a small amount of volatile organic compounds in the flue gas will both cause the flue gas to heat up, the total heat released by the catalytic oxidation reaction is therefore... It can be represented as: in, The calorific value of carbon monoxide. ; This indicates the heat generated by organic matter during catalytic oxidation reactions.
[0031] Therefore, the temperature rise in flue gas caused by the oxidation of carbon monoxide and a small amount of volatile organic compounds can be expressed as: in, For the density of the flue gas, ; Specific heat capacity of flue gas, unit: , , This indicates the composition of the flue gas at flue gas temperature T. Specific heat capacity, Indicates the components in flue gas Volume fraction.
[0032] However, since the concentration of volatile organic compounds is usually low, their exothermic effect is small, and the related measurement equipment is expensive. Therefore, in this embodiment, the exothermic oxidation of volatile organic compounds is not included as an independent input in the mechanism calculation. The small temperature deviation caused by the exothermic reaction of volatile organic compounds is corrected by the time-series residual compensation of the subsequent LSTM model, so as to reduce the control cost.
[0033] Therefore, in this embodiment, the inlet carbon monoxide concentration is considered. and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Represented as: in, The calorific value of carbon monoxide. ; For the density of the flue gas, ; Specific heat capacity of flue gas, unit: .
[0034] Obtaining the initial flue gas temperature Target temperature at the inlet of the catalytic reactor Temperature rise in smoke Based on this, the gas supplementary heating requirement can be calculated. : Assume the gas volume is Gas supplemental heating demand With gas volume The relationship between them is represented as follows: When gas heating demand When the value is greater than 0, the required amount of gas is further obtained. Represented as: in, The calorific value of the gas. Gas volume The unit is .
[0035] Furthermore, the LightGBM model (Lightweight Gradient Boosting Machine) in this embodiment uses the current flue gas volume... Inlet carbon monoxide concentration Initial flue gas temperature ,airspeed and gas heating demand Given the input features, output the current gas reference quantity, in units of... .
[0036] Specifically, the LightGBM model is trained offline based on historical operating data, learning the complex nonlinear mapping relationship between the input features and gas demand. It automatically filters key influencing factors and anchors static mapping patterns through feature importance assessment, filtering out redundant and interfering features to ensure stable and drift-free prediction results. For steady-state and slowly varying operating conditions, it outputs a high-precision gas baseline quantity to meet the precise adjustment requirements under stable operating conditions. It also possesses rapid inference capabilities, adapting to real-time control requirements in industrial settings, and the model output has clear physical meaning, facilitating debugging and understanding by on-site maintenance personnel.
[0037] Furthermore, in this embodiment, the LSTM (Long Short-Term Memory) model takes the historical gas prediction error sequence of the LightGBM model as input and outputs the temporal residual compensation value at the current moment. The historical gas prediction error sequence consists of the predicted gas errors from several past sampling moments. The predicted gas error at each sampling moment is equal to the difference between the actual optimal gas quantity at that moment and the baseline gas quantity predicted by the LightGBM model. This difference includes the temperature rise error caused by the oxidation of volatile organic compounds. This embodiment compensates for the temperature rise error caused by volatile organic compounds by relying on temporal regularity, thus avoiding the need for expensive equipment to collect volatile organic compound-related parameters.
[0038] Specifically, the actual optimal gas flow rate is obtained by back-calculation using feedback from the catalytic reactor inlet temperature: when the catalytic reactor inlet temperature... Deviation from the target temperature at the catalytic reactor inlet At that time, the gas volume difference is calculated by reverse calculation based on the degree of deviation (i.e., the difference between the two). The difference in gas volume The actual optimal gas quantity can be obtained by summing the final gas quantity obtained from the LightGBM model and the LSTM model.
[0039] Furthermore, based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor Back-calculation of gas volume difference Represented as: This embodiment utilizes its long short-term memory network structure to capture timing characteristics such as thermal inertia hysteresis, sudden changes in flue gas flow, and disturbance accumulation, predicting error development trends in advance and outputting timing residual compensation values. This embodiment suppresses prediction drift in this way, enabling control actions to occur before temperature changes, truly achieving proactive adjustment and overcoming the hysteresis problem of traditional control at its root.
[0040] In this embodiment, the current gas reference quantity output by the LightGBM model and the current time-series residual compensation value output by the LSTM model are directly superimposed to obtain the final gas quantity at the current moment. This superposition method is additive fusion, that is, the final gas quantity = gas reference quantity + time-series residual compensation value. The final gas quantity is used as a feedforward control command to adjust the gas valve opening in advance, so as to achieve a proactive match between heat demand and operating condition changes and suppress temperature fluctuations from the source.
[0041] Furthermore, when When the temperature is ≤ 0, the gas heating is turned off, and the temperature is maintained by the self-heating of carbon monoxide and volatile organic compounds; in subsequent cycles if If the value returns to positive, gas heating will resume.
[0042] Furthermore, when the current operating conditions are not within the stable control range, an adaptive adjustment process for the catalytic reactor inlet temperature is executed, specifically: Real-time acquisition of catalytic reactor inlet temperature Based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor The difference between the values is used to calculate the gas adjustment amount. If the gas adjustment amount is positive, the gas amount is increased; if the gas adjustment amount is negative, the gas amount is decreased.
[0043] Specifically, based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor The difference between them is used to calculate the gas adjustment amount, expressed as: Among them, if If positive, then increase the final gas volume based on the previous moment. The amount of gas, if If the value is negative, then the final gas volume will be reduced from the previous moment's value. The amount of gas.
[0044] This embodiment integrates the physical mechanism of space velocity-conversion rate with LightGBM-LSTM intelligent prediction to achieve proactive regulation and precise control of flue gas heating. Compared with the traditional hysteresis feedback regulation method that relies on outlet temperature deviation, this embodiment predicts the contribution of carbon monoxide self-exothermic emission to flue gas temperature in advance through mechanism prediction, enabling control commands to be issued in advance of temperature changes. This effectively suppresses temperature fluctuations caused by thermal inertia and avoids instantaneous excessive emissions of carbon monoxide. LightGBM is responsible for high-precision benchmark fitting under steady-state conditions, while LSTM is specifically responsible for time-series residual compensation caused by thermal inertia hysteresis, sudden changes in operating conditions, and exothermic emissions from volatile organic compounds. The two have clear division of labor and complement each other, ensuring prediction accuracy while eliminating the need for expensive equipment to collect volatile organic compound concentration data, thus reducing system implementation costs. At the same time, by implementing a dual-mode strategy of mechanism-driven proactive regulation in the stable range and switching to outlet temperature adaptive regulation in the unstable range, stable temperature control is achieved across the entire operating range, ensuring energy-saving operation under normal conditions and ensuring safety under extreme conditions.
[0045] Example 2: This embodiment provides a flue gas heating advance regulation system based on space velocity-conversion rate driven LightGBM-LSTM, including a memory and a processor. The memory stores a computer program, and when the processor runs the computer program, it executes the flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM in Embodiment 1.
[0046] Example 3: This embodiment provides a storage medium storing a computer program, which is executed by a processor to perform the flue gas heating advance regulation method based on airspeed-conversion rate driven LightGBM-LSTM in Embodiment 1.
[0047] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for advance regulation of flue gas heating based on space velocity-conversion rate driven LightGBM-LSTM, characterized in that, include: The operating parameters of the sintering flue gas heating system are collected, including the flue gas volume. Inlet carbon monoxide concentration Initial flue gas temperature and catalytic reactor inlet temperature ; The collected operating parameters are preprocessed and adjusted according to the flue gas volume. With catalyst volume Calculate airspeed ; According to airspeed and inlet carbon monoxide concentration Determine whether the current operating condition is within a stable control range; If the system is within a stable control range, then the mechanism-driven anti-adjustment process will be executed: Based on airspeed Calculate the current carbon monoxide conversion rate And based on the inlet carbon monoxide concentration and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Based on the initial flue gas temperature Target temperature at the inlet of the catalytic reactor Temperature rise in smoke Calculate the gas supplementary heating demand ; like If the value is greater than 0, the current gas reference quantity is output through the LightGBM model, and the current time-series residual compensation value is output through the LSTM model. The final gas quantity at the current moment is obtained by superimposing the gas reference quantity and the time-series residual compensation value. If... If the value is ≤ 0, then the gas heating will be turned off; If the temperature is not within the stable control range, then the adaptive adjustment process for the catalytic reactor inlet temperature will be executed: Based on the inlet temperature of the catalytic reactor With the target temperature at the inlet of the catalytic reactor The deviation is used to calculate the gas adjustment amount, and the final gas amount at the current moment is updated to be the sum of the final gas amount at the previous moment and the gas adjustment amount.
2. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, According to the amount of flue gas With catalyst volume Calculate airspeed Represented as: Among them, catalyst volume The catalyst loading volume in the catalytic reactor; space velocity. This represents the gas volume flow rate passing through a unit volume of catalyst per unit time.
3. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, If the current inlet carbon monoxide concentration In the range of 3 to 10 Within range and airspeed Between 9000 and 40000 If the current condition is within the specified range, it is considered to be in a stable control range; otherwise, it is considered not to be in a stable control range.
4. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, Based on the inlet carbon monoxide concentration and carbon monoxide conversion rate Calculate the flue gas temperature rise caused by the exothermic oxidation of carbon monoxide. Represented as: in, The calorific value of carbon monoxide; The density of the flue gas; This refers to the specific heat capacity of the flue gas.
5. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, The LightGBM model uses the current amount of smoke. Inlet carbon monoxide concentration Initial flue gas temperature ,airspeed and gas heating demand Given the input features, output the current gas reference quantity.
6. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, The LSTM model takes the historical gas prediction error sequence of the LightGBM model as input and outputs the time-series residual compensation value at the current moment. Among them, the historical predicted gas error sequence consists of the predicted gas error at several past sampling times. The predicted gas error at each sampling time is equal to the difference between the actual optimal gas quantity at that time and the gas reference quantity predicted by the LightGBM model. This difference includes the temperature rise error caused by the oxidation of volatile organic compounds.
7. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, The amount of gas regulation in the adaptive temperature control process of the catalytic reactor inlet is expressed as: in, The density of the flue gas; The specific heat capacity of the flue gas; The calorific value of the gas; For gas regulation, if If positive, then increase the final gas volume based on the previous moment. The amount of gas, if If the value is negative, then the final gas volume will be reduced from the previous moment's value. The amount of gas.
8. The flue gas heating advance regulation method based on space velocity-conversion rate driven LightGBM-LSTM according to claim 1, characterized in that, The gas heating demand Equal to the target temperature at the inlet of the catalytic reactor Subtract the initial flue gas temperature Temperature rise in smoke .
9. A flue gas heating advance regulation system based on space velocity-conversion rate driven LightGBM-LSTM, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the flue gas heating advance regulation method based on airspeed-conversion rate driven LightGBM-LSTM as described in any one of claims 1-8 when running the computer program.
10. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, performs the flue gas heating advance regulation method based on airspeed-conversion rate driven LightGBM-LSTM as described in any one of claims 1-8.