Intelligent Control Method for Boiler Spraying Based on Closed-Loop Feedback of Flue Gas Temperature

By adopting an intelligent control method based on closed-loop feedback of flue gas temperature, the control stability problem of boiler spray system under frequent load changes and variable combustion conditions is solved, realizing refined temperature control and improved automation, and reducing operation and maintenance difficulty.

CN122943334BActive Publication Date: 2026-11-17CHANGCHUN NO 2 CO GENERATION POWER
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611231216.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-08-14
Publication Date
2026-11-17
Estimated Expiration
2046-08-14

AI Technical Summary

Technical Problem

Existing boiler spray systems suffer from poor control stability and slow response under frequent load changes and varying combustion conditions. They are prone to overheating and oscillations, making it difficult to achieve precise temperature control. They rely on manual intervention and have a low degree of automation.

Method used

An intelligent control method based on closed-loop feedback of flue gas temperature is adopted. By dimensionlessly normalizing the flue gas temperature and spray valve position, combined with adaptive PID control, feedforward compensation and inverse function nonlinear compensation, and using oxygen content deviation prediction and trend analysis, the spray command can be accurately adjusted, and the control parameters are updated adaptively through statistical methods.

Benefits of technology

It improves the automatic operation rate of the spray system at full load, reduces low-load stagnation and high-load oscillation, enhances the response capability to combustion disturbances, avoids overheating of the denitrification system, and reduces the difficulty of operation and maintenance and the risk of unplanned shutdowns.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122943334B_ABST
    Figure CN122943334B_ABST
Patent Text Reader

Abstract

This invention relates to the field of boiler automation control and discloses an intelligent boiler spray control method based on closed-loop feedback of flue gas temperature, applicable to industrial boilers with frequent load changes and variable combustion conditions. The method includes dimensionless normalization of real-time collected flue gas temperature and valve position; determining adaptive control parameters through linear interpolation based on the dimensionless temperature values, performing closed-loop calculations based on temperature deviations, and then superimposing adaptive feedforward compensation based on flue gas oxygen content deviations; using a field-calibrated inverse function to eliminate dead zones and correct nonlinearities in the superimposed commands, while simultaneously using a logarithmic formula to predict increments based on temperature change rate trends; and finally, outputting the results to the spray actuator after integral anti-saturation and rate limiting. This invention does not rely on complex machine learning models, effectively overcoming valve dead zones and large system delays, and significantly improving the dynamic response speed, steady-state accuracy, and adaptive capability of temperature control.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of boiler automation control, and in particular to a method for intelligent control of boiler spraying based on closed-loop feedback of flue gas temperature. Background Technology

[0002] In the operation system of thermal power plants and industrial boilers, precise control of flue gas temperature in the tail flue is a core process to ensure the safety of denitrification catalysts (preventing ammonia escape and crystallization or catalyst sintering), reduce flue gas heat loss, and meet ultra-low emission requirements. Currently, the industry generally uses spray desuperheating water to level the flue gas temperature, and the control logic is mostly based on single-loop PID closed loop. Some newly built units have introduced cascade control or simplified Smith predictor compensation.

[0003] However, as deep peak shaving has become the norm in my country's power system, boilers are frequently subjected to wide load ranges, with inconsistent coal blending ratios and dynamic adjustments to air distribution strategies. This results in highly nonlinear and large-hysteresis characteristics in the thermal performance of the flue gas at the tail end. Against this backdrop, the limitations of traditional temperature control logic in actual operation have become increasingly prominent. On-site operations still heavily rely on manual intervention and experience-based tuning by operators, and the stability of automatic operation of sprinkler systems is generally low.

[0004] Based on current industrial practices, existing technologies mainly suffer from the following pain points: The PID parameters were fixed during the commissioning phase and did not migrate with operating conditions. This resulted in a sluggish response at low loads, leading to overheating, and excessively rapid integral accumulation at high loads, causing significant fluctuations in water spray volume and even triggering the denitrification system to switch to manual mode.

[0005] Using only the flue gas temperature after spraying as the sole controlled variable is a typical example of reactive regulation. When there are sudden changes in the combustion side air volume or coal quality, the oxygen content change has already occurred for several minutes before the temperature shows up, meaning the control is always a step behind.

[0006] The controller only performs algebraic calculations based on the current deviation and the rate of change of deviation, without any trend prediction logic. When encountering a sudden increase in flue gas temperature (such as soot blowing disturbance), the PID controller has not yet fully responded before the temperature exceeds the environmental protection limit.

[0007] Model predictive control (MPC) and neural network algorithms proposed in recent years rely on accurate heat transfer models or massive amounts of labeled data. Industrial field calibration is extremely costly, and black-box algorithms have weak interpretability when abnormal operating conditions occur. Power plants rarely put automation into long-term operation due to safety concerns.

[0008] Therefore, we propose a boiler spray intelligent control method based on flue gas temperature closed-loop feedback to solve the above problems. Summary of the Invention

[0009] This invention provides a boiler spray intelligent control method based on flue gas temperature closed-loop feedback, which is used for industrial boiler spray cooling scenarios with frequent load changes and varied combustion conditions.

[0010] The first aspect of this invention provides a method for intelligent boiler spray control based on closed-loop feedback of flue gas temperature. The method includes: performing dimensionless normalization on the collected flue gas temperature and spray valve position to obtain dimensionless temperature values ​​and dimensionless valve position values; determining adaptive control parameters based on the dimensionless temperature values ​​and their rate of change; and performing closed-loop control calculations based on the temperature deviation between the dimensionless temperature values ​​and the set temperature to obtain an initial spray command; calculating a feedforward compensation amount based on the oxygen content deviation between the flue gas oxygen content and the set oxygen content, and then applying the feedforward compensation to the set oxygen content. The feedforward compensation amount is combined with the initial spray command to obtain a superimposed spray command; the superimposed spray command is nonlinearly compensated using a pre-established inverse function to obtain a compensated spray command; then, a trend prediction is made based on the rate of change of the dimensionless temperature value, and the predicted increment is superimposed on the compensated spray command to obtain an adjusted spray command; the adjusted spray command is restricted and then output to the spray actuator; simultaneously, based on statistical methods, the continuous operation data is statistically analyzed to adaptively update the compensation parameters of the inverse function and the amplitude of the predicted increment.

[0011] Optionally, in a first implementation of the first aspect of the present invention, the method includes: The difference between the measured value of the oxygen content in the flue gas and the set value of the oxygen content is calculated as the oxygen content deviation; The oxygen content deviation is multiplied by a feedforward coefficient to obtain the feedforward compensation amount, wherein the feedforward coefficient is adaptively adjusted by statistically analyzing the sign relationship between the temperature deviation and the oxygen content deviation online. The feedforward compensation amount is added to the initial spray command to obtain the superimposed spray command.

[0012] Optionally, in a second implementation of the first aspect of the present invention, the adaptive adjustment process of the feedforward coefficient includes: The oxygen content deviation and the temperature deviation are collected over multiple consecutive control cycles. If the oxygen content deviation and the temperature deviation have the same sign, then the current feedforward coefficient is increased by a preset step increment. If the oxygen content deviation has the opposite sign to the temperature deviation, then the current feedforward coefficient is reduced by a preset step decrease. Wherein, both the step increment and the step decrement are fixed values.

[0013] Optionally, in a third implementation of the first aspect of the present invention, the method includes: The superimposed spray command is input into the inverse function, and the valve dead zone is eliminated and the gain nonlinearity is corrected through the inverse function to obtain the compensated spray command. When the rate of change of the dimensionless temperature value is detected to be rapidly increasing and the dimensionless temperature value exceeds a preset high-temperature threshold, the predicted increment is added to the compensated spray command; when the rate of change of the dimensionless temperature value is detected to be rapidly decreasing and the dimensionless temperature value is lower than a preset low-temperature threshold, the predicted increment is reduced to the compensated spray command; otherwise, the compensated spray command remains unchanged; the above adjusted result is used as the adjusted spray command.

[0014] Optionally, in the fourth implementation of the first aspect of the present invention, the inverse function is obtained through on-site calibration: the actual flow rate value corresponding to the three different opening points of the spray actuator is measured respectively, and three fitting coefficients are obtained by fitting a quadratic curve using the correspondence between the three opening points and the flow rate value, and the fitting coefficients constitute the compensation parameters of the inverse function. The rapid rise refers to the rate of change of the dimensionless temperature value exceeding a preset rise rate threshold for three consecutive sampling cycles, and the rapid fall refers to the rate of change of the dimensionless temperature value being lower than a preset fall rate threshold for three consecutive sampling cycles. The magnitude of the predicted increment is positively correlated with the absolute value of the rate of change of the dimensionless temperature value, and the predicted increment increases or decreases with a fixed step size until the trend is eliminated and returns to zero.

[0015] Optionally, in a fifth implementation of the first aspect of the present invention, the predicted increment is calculated using the following logarithmic formula:

[0016] in, To predict the increment; This represents the rate of change of the dimensionless temperature value during the current sampling period. This is a preset operating condition ratio coefficient; is the preset rate amplification factor; m is the count value for continuous trigger trend judgment; This is the step value with a fixed step size.

[0017] Optionally, in a sixth implementation of the first aspect of the present invention, the method includes: The adjusted spray command is subjected to integral anti-saturation processing: it is determined whether the adjusted spray command exceeds the preset upper or lower limit. If it exceeds the upper or lower limit, only the proportional and derivative parts in the closed-loop control calculation are retained, and integral accumulation is restored when the command returns to the range between the upper and lower limits to obtain the anti-saturation processed spray command. The spray command after the anti-saturation treatment is rate-limited: the spray command after the anti-saturation treatment is compared with the spray command of the previous control cycle. If the difference between the two exceeds the preset maximum allowable change rate, the current output is limited to the spray command of the previous control cycle plus or minus the maximum allowable change rate, so as to obtain the spray command output to the spray actuator.

[0018] Optionally, in a seventh implementation of the first aspect of the present invention, the method includes: Using a sliding window statistical method, temperature overshoot and valve command deviation are collected for multiple consecutive operating cycles. The median or mean of the temperature overshoot and the median or mean of the valve command deviation are calculated respectively. The amplitude of the predicted increment is adaptively adjusted based on the statistical results of the temperature overshoot. The compensation parameters of the inverse function are adaptively adjusted based on the statistical results of the valve command deviation. The updated compensation parameters and amplitudes are used for the adjustment processing in the next operating cycle.

[0019] Optionally, in an eighth implementation of the first aspect of the present invention, an initialization calibration process is also included: When the system starts up or when the boiler load change exceeds a preset threshold, the spray valve is forced to a preset reference opening, and the stable temperature value under the stable state is recorded after the flue gas temperature reaches a stable state. Based on the difference between the stable temperature value and the design reference temperature value, the zero-point offset of the inverse function is calculated and updated to obtain the calibrated inverse function compensation parameters, which are used for subsequent nonlinear compensation.

[0020] Optionally, in a ninth implementation of the first aspect of the present invention, the reference opening is the middle position of the entire stroke of the spray valve; The condition for determining the stable state is that the rate of change of the dimensionless temperature value is continuously lower than a preset stability threshold within multiple consecutive sampling periods. The zero-point offset is calculated as follows: the difference between the stable temperature value and the design reference temperature value is multiplied by a preset correction coefficient to obtain the offset value; the offset value is then added to the constant term of the inverse function to replace the original compensation parameter.

[0021] The mechanism of this invention is as follows: It organically unifies the robustness of PID control, the speed of feedforward control, the accuracy of inverse function compensation, and the adaptability of statistical methods, effectively balancing control accuracy, response speed, anti-interference ability, and engineering maintainability without introducing complex intelligent algorithms and online modeling. Beneficial effects: By combining dimensionless processing with piecewise linear interpolation, the PID parameters change smoothly and gradually with the temperature range. This eliminates gain jumps during operating condition switching and eliminates the need to introduce complex membership rules of fuzzy control. The automatic control operation rate of the sprinkler system is significantly improved across the entire load range, and the industry common problems of low-load stagnation and high-load oscillation are eliminated. By utilizing the consistency of the sign between the oxygen content deviation and the control deviation, step-by-step online parameter adjustment is performed. Combined with the valve inverse function based on three-point calibration, nonlinear physical compensation and combustion disturbance offsetting are completed before the command is issued. When there are sudden changes in coal quality or air distribution adjustment, the spray response is advanced. Instead of simple deviation differential calculation, it superimposes the directional prediction increment based on the slope trend judgment of the dimensionless temperature continuous sampling period. The logic is clear and there is no risk of noise amplification. Without introducing filtering delay, it successfully suppresses the sharp rise and fall of flue gas temperature and avoids environmental exceedance incidents caused by excessive temperature at the denitrification inlet. The system completes the self-correction of inverse function offset and predicted amplitude by relying solely on the median / mean statistics of continuous running cycles, achieving calibration while running. The system has a self-learning capability while retaining white-box logic, allowing maintenance personnel to intuitively trace the basis for each parameter adjustment, greatly reducing the resistance to promotion. After the dead zone compensation and prediction are superimposed, a conditional anti-saturation recovery mechanism and pre- and post-cycle rate clamping are applied. This not only eliminates the frequent chattering of the spray valve in the small deviation zone, but also substantially extends the maintenance cycle of the electric / pneumatic actuator and positioner, reducing unplanned downtime. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of an embodiment of the intelligent control method for boiler spraying based on closed-loop feedback of flue gas temperature in this invention. Figure 2 This is a schematic diagram of another embodiment of the intelligent control method for boiler spraying based on closed-loop feedback of flue gas temperature in this invention. Figure 3 This is a schematic diagram showing how the original flue gas temperature and spray valve position data are converted into dimensionless values ​​between 0 and 1. The flue gas temperature on the left and the spray valve position on the right both have an original measurement range. Figure 4 This is a schematic diagram of an embodiment of the intelligent boiler spray control device based on closed-loop feedback of flue gas temperature in this invention. Detailed Implementation

[0023] This invention provides an intelligent boiler spray control method based on closed-loop feedback of flue gas temperature, applicable to industrial boilers with frequent load changes and variable combustion conditions. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the intelligent boiler spray control method based on closed-loop feedback of flue gas temperature in this invention includes: 101. Perform dimensionless normalization on the real-time collected flue gas temperature and spray valve position to obtain dimensionless temperature value and dimensionless valve position value.

[0025] It is understood that the executing entity of this invention can be a boiler spray intelligent control device based on flue gas temperature closed-loop feedback, or it can be a terminal or a server; the specific implementation is not limited here. This embodiment of the invention will be described using a server as an example.

[0026] It should be noted that the specific execution process of this step is illustrated using the quench spray tower control system of a municipal solid waste incineration boiler as an example. During the system initialization phase, based on the boiler's quench tower operating specifications, personnel set the effective monitoring lower limit for flue gas temperature in the distributed control system to 150℃ and the effective monitoring upper limit to 250℃. Therefore, the temperature range under this process environment is 100℃. Simultaneously, the regulating valve controlling the spray water flow rate uses a standard control signal, with a set lower limit of 0% and an upper limit of 100%, resulting in a valve range of 100%.

[0027] During a sampling cycle in the boiler operation, the temperature sensor installed at the outlet of the quench tower transmits the currently measured real physical data to the control system, displaying the current actual flue gas temperature as 212℃. At the same time, the intelligent valve positioner on the spray pipeline transmits its real-time status, displaying the current actual valve opening as 45%. The system immediately normalizes these two data points.

[0028] For flue gas temperature processing: The system subtracts the set lower limit of 150℃ from the currently collected 212℃, calculating a difference of 62℃. Then, the system divides this difference by the set temperature range span of 100℃. This division gives the dimensionless temperature value for the current cycle as 0.62. For spray valve position processing: The system subtracts the lower limit of 0% from the currently collected 45% opening, resulting in a difference of 45%. This difference is then divided by the valve position range span of 100%. This calculation gives the dimensionless valve position value for the current cycle as 0.45. Through these single-step operations, the control system converts the raw data into pure digital variables uniformly within the 0~1 range.

[0029] 102. Divide the working sections according to the dimensionless temperature value and its rate of change, determine the adaptive PID parameters of the sections by linear interpolation, and perform PID calculation based on the deviation between the dimensionless temperature value and the set value to obtain the initial spray command.

[0030] It should be noted that the process requires the flue gas temperature to be stably controlled at 200℃. The system normalizes this target value: subtracting 150 from 200 and then dividing by 100, yields a dimensionless setpoint of 0.50. Entering the current sampling cycle, the control system receives a current dimensionless temperature value of 0.62, and retrieves the dimensionless temperature data from the previous sampling cycle, which is 0.60. Comparison shows that the current dimensionless temperature change rate is +0.02, indicating that the system is in a continuous heating state.

[0031] The system consults its internal matrix and determines that the current operating condition has entered the high-temperature rising phase. In this phase's baseline, a dimensionless temperature of 0.60 corresponds to a proportional gain of 1.5 and an integral gain of 0.20; 0.70 corresponds to a proportional gain of 2.5 and an integral gain of 0.30. The system performs linear interpolation: 0.62 lies at 1 / 5 of the interval between these two values. Using the same proportional conversion, the adaptive proportional gain for the current cycle is calculated to be 1.7, the integral gain to be 0.22, and the derivative gain to be 0.06.

[0032] After the parameters are determined, the system calculates the control deviation by subtracting the target value of 0.50 from the current temperature of 0.62, resulting in a positive deviation of 0.12. The control unit amplifies this deviation using a proportional gain of 1.7, calculating the proportional control action output for this cycle to be 0.204. Simultaneously, the system retrieves the integral action output value generated from historical accumulated deviations, currently recorded as 0.376. The system adds the 0.204 generated by the proportional action to the historical accumulated integral action output value of 0.376, and combines this with the slight differential action to obtain an initial spray command with a value of 0.58.

[0033] 103. Calculate the feedforward compensation amount based on the deviation of oxygen content in flue gas, and superimpose the feedforward compensation amount with the initial spray command to obtain the superimposed spray command.

[0034] It should be noted that the initial spray command, calculated based on pure temperature feedback, was 0.58. In this step, the baseline target value for the oxygen content in the boiler flue gas was set at 7.0%. At the current sampling time, the real-time oxygen content data transmitted from the end-of-flue sensor was 5.5%.

[0035] The system performs deviation calculations: subtracting the actual oxygen content of 5.5% from the target oxygen content of 7.0% yields a current negative deviation of 1.5%. A rapid decrease in flue gas oxygen content usually indicates a sudden increase in the calorific value of the waste entering the furnace, foreshadowing a subsequent surge of high-temperature hot air into the quench tower. To cope with this impending thermal shock, the control system activates feedforward compensation. The pre-calibrated feedforward gain rule within the system is set as follows: for every 1% drop in actual oxygen content below the target value, the dimensionless spray command must be increased by 0.03 in advance to accelerate cooling.

[0036] The current 1.5% hypoxia deviation is multiplied by a coefficient of 0.03, resulting in a current feedforward compensation of 0.045. The system then executes a command superposition operation. The initial spray command of 0.58 generated from temperature feedback is added to the feedforward compensation of 0.045 generated from oxygen content prediction, yielding a result of 0.625, which is then used as the superimposed spray command.

[0037] 104. Use the pre-calibrated inverse function to perform valve dead zone and nonlinear compensation on the superimposed spray command to obtain the compensated spray command; then, based on the trend of the change rate of the dimensionless temperature value, when the temperature rises rapidly and exceeds the high temperature threshold or falls rapidly and is below the low temperature threshold, add a predicted increment to the compensated spray command to obtain the adjusted spray command.

[0038] It should be noted that valve physical characteristic compensation was performed. Based on calibration data during maintenance, the valve exhibited a mechanical dead zone and non-linear flow. The system substituted 0.625 into its internal inverse function mapping table. The table data showed that, to accurately generate the expected flow rate, an opening command of 0.660 was actually required for the valve. Therefore, the system obtained a compensated spray command of 0.660.

[0039] Entering the trend prediction stage, the dimensionless high-temperature alarm threshold set for this quench tower is 0.60, and the rapid change criterion for the dimensionless temperature change rate is ±0.015. The system checks the current data: the dimensionless temperature value for the current period is 0.62, which is greater than the set high-temperature threshold of 0.60; at the same time, the dimensionless temperature change rate for the current period is +0.02, which is also greater than the rapid rise criterion of 0.015. The simultaneous fulfillment of these two conditions means that the tower is experiencing a rapid and drastic temperature increase.

[0040] According to process safety specifications, when this condition is triggered, a pre-judgment increment of 0.04 must be added to forcibly increase the water spray volume. The system adds the compensated spray command of 0.660 to the pre-judgment increment of 0.04. After this calculation, the system outputs an adjusted spray command of 0.70, which, while addressing valve nonlinearity, also includes pre-judgment protection.

[0041] 105. Perform integral anti-saturation processing and rate limiting processing on the adjusted spray command to obtain the spray command output to the actuator; at the same time, based on the sliding window statistical method, adaptively update the compensation parameters of the inverse function and the magnitude of the predicted increment. The statistical method uses the running data of multiple consecutive cycles to calculate the median or mean, without relying on the machine learning model.

[0042] It should be noted that, due to equipment protection constraints and process safety regulations, the maximum opening change rate of this valve within a single sampling period must not exceed 0.03. The system retrieves the output command from the previous period and finds its value to be 0.65. Since the difference between the current command 0.70 and the previous command 0.65 is +0.05, exceeding the action limit of 0.03, the system immediately triggers the rate limit, forcibly truncating the opening increase of this period to 0.03. The calculated spray command is 0.68. This value is also within the valid range, allowing the integral anti-saturation module to operate normally in the background. The system converts 0.68 into an electrical signal output, driving the valve to open to 68%.

[0043] Simultaneously with the issuance of the command, a background sliding window recorded valve data from the most recent five periods for median calculation. The relevant data records are shown in Table 1 below: Table 1

[0044] Because the system's issued commands are in a continuous and stable upward trend, the on-site feedback opening data also shows a synchronous upward trend, conforming to the physical law of following. The system sorts the steady-state positioning deviation values ​​of these 5 cycles (1.3%, 1.4%, 1.5%, 1.5%, 1.6%) and extracts the value in the middle, 1.5% (i.e., 0.015). Based on this, the system determines that the valve has a fixed steady-state hysteresis deviation due to wear and automatically adjusts the compensation benchmark parameter of the inverse function in step 104 by 0.015. The system utilizes this low-computing-power pure statistical mechanism to complete the adaptive fine-tuning of the system's long-term operating parameters.

[0045] Please see Figure 2 and Figure 3 Another embodiment of the intelligent boiler spray control method based on closed-loop feedback of flue gas temperature in this invention includes: 201. Perform dimensionless normalization on the real-time collected flue gas temperature and spray valve position to obtain dimensionless temperature value and dimensionless valve position value.

[0046] Specifically, the lower and upper limits of the flue gas temperature allowed by the boiler design, as well as the lower and upper limits of the spray valve position stroke, are obtained. Then, the lower limit of the temperature is subtracted from the real-time measured flue gas temperature, and then divided by the difference between the upper and lower limits of the temperature to obtain the dimensionless temperature value. At the same time, the lower limit of the valve position stroke is subtracted from the real-time measured spray valve position, and then divided by the difference between the upper and lower limits of the valve position stroke to obtain the dimensionless valve position value. The dimensionless temperature value and the dimensionless valve position value are both limited to the range of 0 to 1 and are used as inputs for subsequent processing.

[0047] It should be noted that the quenching spray system of a hazardous waste incineration boiler in a chemical industrial park is used as an example to illustrate the specific execution logic of this step. During the configuration phase before system startup, the process engineer sets the lower limit of flue gas temperature to 180℃ and the upper limit of flue gas temperature to 280℃, based on the safety specifications of the boiler's quenching process, thus deriving a temperature range of 100℃.

[0048] The lower limit of the stroke of the pneumatic regulating valve for adjusting the spray water volume is set to 0% (representing the valve is fully closed), and the upper limit of the stroke is set to 100% (representing the valve is fully open). The range of the valve position is 100%. These reference parameters are fixedly written into the control system as the basis for normalization.

[0049] During a certain control cycle under normal boiler operation, the real-time flue gas temperature collected by the temperature sensor installed at the outlet of the quench tower is 225℃, while the real-time spray valve position reported by the valve positioner is 35%. The system performs dimensionless calculation of the flue gas temperature: subtracting the lower limit of 180℃ from the real-time collected 225℃, the calculated difference is 45℃.

[0050] Dividing 45℃ by the temperature range of 100℃, the initial result is calculated to be 0.45. The system checks the result and confirms that 0.45 is within the range of 0 to 1, so it is deemed valid and the dimensionless temperature value output is 0.45.

[0051] Similarly, the system calculates the spray valve position: subtracting the lower limit of the stroke (0%) from the real-time collected 35% opening, the difference is 35%. Then, 35% is divided by the valve position range span of 100%, yielding an initial result of 0.35. The system verifies that this value is also within the valid range, and outputs a dimensionless valve position value of 0.35.

[0052] To demonstrate the range limiting function, it is assumed that the measured temperature collected by the sensor reaches 295℃ during the subsequent abnormal fluctuation period. The system subtracts the lower limit of 180℃ from 295℃, obtaining a difference of 115℃; after dividing this by the range of 100℃, the initial result is 1.15.

[0053] The interval restriction processing mechanism was triggered. The system recognized that 1.15 exceeded the upper limit of 1, so it automatically truncated it. The dimensionless temperature value output was restricted to 1. The original physical measurement point data on site was successfully transformed into a standardized, constrained pure digital variable of 0 to 1.

[0054] 202. Divide the working sections according to the dimensionless temperature value and its rate of change, determine the adaptive PID parameters of the sections by linear interpolation, and perform PID calculation based on the deviation between the dimensionless temperature value and the set value to obtain the initial spray command.

[0055] Specifically, based on the numerical range of the dimensionless temperature value, the operating conditions are divided into low-temperature, medium-temperature, and high-temperature zones. Each zone has a preset proportional gain base value, integral gain base value, and derivative gain base value. Then, based on the specific position of the dimensionless temperature value within the current zone, the adaptive PID parameters for the current zone are calculated between the base values ​​of adjacent zones using linear interpolation. These parameters include the proportional gain coefficient, integral gain coefficient, and derivative gain coefficient. The difference between the dimensionless temperature value and the set value is used as a deviation signal. The adaptive PID parameters for the zone are used to perform proportional, integral, and derivative operations on this deviation signal to obtain the initial spray command. This initial spray command is used for subsequent superposition processing.

[0056] Furthermore, the boundary of the low-temperature zone is a dimensionless temperature value less than 0.35, the boundary of the medium-temperature zone is a dimensionless temperature value between 0.35 and 0.65, and the boundary of the high-temperature zone is a dimensionless temperature value greater than 0.65. Linear interpolation uses the base values ​​of adjacent zones for weighted calculation. The weight is determined by the relative position of the dimensionless temperature value within the current zone. That is, when the current zone is close to the previous zone, the base value of the previous zone is mainly used, and when it is close to the next zone, the base value of the next zone is mainly used, thus obtaining continuously changing zone adaptive PID parameters. The initial spray command is the result obtained after proportional, integral, and derivative operations, and this result is used as the input for subsequent processing.

[0057] It should be noted that, continuing with the aforementioned scenario of the rapid cooling spray system for hazardous waste incineration boilers in chemical industrial parks, the process requires the flue gas temperature to be stably controlled at 226℃. Based on the range conversion of the preceding steps, the dimensionless setpoint for this target is 0.460. Entering the current sampling cycle, the control system receives a current dimensionless temperature value of 0.560. Since 0.560 falls between 0.350 and 0.650, the system determines that the current operating condition is in the medium-temperature range. Simultaneously, the system retrieves the dimensionless temperature change rate for the current cycle, which is positive 0.020.

[0058] The adaptive parameter linear interpolation calculation begins. The anchor point for the mid-temperature range is set at 0.500, with a proportional gain base value of 2.0, an integral gain base value of 0.20, and a differential gain base value of 0.05. The anchor point for the high-temperature range is set at 0.800, with a proportional gain base value of 3.0, an integral gain base value of 0.30, and a differential gain base value of 0.10. The current temperature value of 0.560 lies between the two anchor points, 0.060 from the mid-temperature anchor point and 0.240 from the high-temperature anchor point. According to the distance weighting principle, the system assigns 80% weight to the mid-temperature base value and 20% weight to the high-temperature base value. The calculated current proportional gain coefficient is 2.20, calculated as 2.0 multiplied by 0.8 for the mid-temperature range and 3.0 multiplied by 0.2 for the high-temperature range. Similarly, the integral gain coefficient is calculated to be 0.22, and the differential gain coefficient is 0.06.

[0059] After the parameters are determined, the system subtracts the target value of 0.460 from the current dimensionless temperature of 0.560, resulting in a positive deviation of 0.100. The control unit amplifies this deviation using a proportional gain coefficient of 2.20, resulting in a proportional control output value of 0.220. The control system retrieves the integral action output value generated by historical cumulative deviations, currently recorded as 0.359. Based on the current temperature change rate of 0.020 and a differential gain coefficient of 0.06, the system calculates the differential output value used to suppress the temperature change trend as 0.0012, which is truncated to three decimal places and recorded as 0.001 within the control system.

[0060] The proportional output of 0.220, the integral output of 0.359, and the derivative output of 0.001 are added together to obtain an initial spray command of 0.580, which represents the expected spray valve opening of 58.0% under the basic feedback logic.

[0061] 203. Calculate the feedforward compensation amount based on the deviation of oxygen content in flue gas, and superimpose the feedforward compensation amount with the initial spray command to obtain the superimposed spray command.

[0062] Specifically, the measured value and the set value of the oxygen content in the flue gas are obtained, and the difference between the two is calculated as the oxygen content deviation. The oxygen content deviation is multiplied by a feedforward coefficient to obtain the feedforward compensation amount. The feedforward coefficient is adaptively adjusted by statistically analyzing the sign relationship between the temperature deviation and the oxygen content deviation online. The feedforward compensation amount is added to the initial spray command to form a superimposed spray command. This superimposed spray command is used for subsequent dead zone and nonlinear compensation and trend prediction adjustment.

[0063] Furthermore, the adaptive adjustment process of the feedforward coefficient includes: collecting data on the direction of change of flue gas oxygen content deviation and deviation signal within multiple consecutive control cycles; if the signs of the oxygen content deviation and deviation signal are the same, increasing the current feedforward coefficient by a preset step increment; if the signs of the oxygen content deviation and deviation signal are opposite, decreasing the current feedforward coefficient by a preset step decrement; both the step increment and the step decrement are fixed values, and the adjusted feedforward coefficient is used to calculate the feedforward compensation amount in the next control cycle, thereby obtaining a new superimposed spray command.

[0064] It should be noted that the system received an initial spray command with a value of 0.580, and the dimensionless temperature deviation signal for this cycle is positive 0.100. The baseline oxygen content of the boiler's flue gas was set at 6.5%. During the current control cycle, due to the influx of high-calorific-value waste liquid into the furnace, a large amount of oxygen is consumed, and the real-time oxygen content of the flue gas measured in the flue has dropped to 4.5%.

[0065] The system calculates the deviation: subtracting the measured value of 4.5% from the target value of 6.5%, the current oxygen content deviation is positive 2.0%. This positive deviation indicates a large amount of high-temperature flue gas will be generated. The system reads the currently effective feedforward coefficient, recording it as 0.020. The system multiplies the oxygen content deviation value of 2.0 by the feedforward coefficient of 0.020, calculating the feedforward compensation for this cycle as 0.040. The system then adds this to the initial spray command of 0.580 from step 202, obtaining 0.620. Thus, the system obtains a 0.620 superimposed spray command containing the intention of advanced cooling, which is then sent to step 204.

[0066] Simultaneously with the instruction stacking, the system initiates adaptive adjustment of the feedforward coefficient. The system checks the data for this period: the oxygen content deviation is +2.0%, and the temperature deviation signal is +0.100. The same sign for both indicates that the temperature rise prediction based on the decrease in oxygen content matches the actual higher-than-expected temperature, confirming the correct direction of the feedforward intervention, which can be enhanced. The system uses a preset step increment of 0.002. Adding this increment to the current feedforward coefficient of 0.020 yields a new feedforward coefficient of 0.022.

[0067] 204. Use the pre-calibrated inverse function to perform valve dead zone and nonlinear compensation on the superimposed spray command to obtain the compensated spray command; then, based on the trend of the change rate of the dimensionless temperature value, when the temperature rises rapidly and exceeds the high temperature threshold or falls rapidly and is below the low temperature threshold, add a predicted increment to the compensated spray command to obtain the adjusted spray command.

[0068] Specifically, an inverse function, pre-calibrated on-site, is obtained. This inverse function maps the ideal dimensionless opening to the actual valve position command. Then, the superimposed spray command is input into the inverse function, and dead zone elimination and gain nonlinearity correction are performed through this inverse function to obtain the compensated spray command. Trend judgment is made based on the dimensionless temperature value and its rate of change: when the rate of change is detected to be rapidly increasing and the dimensionless temperature value exceeds the preset high temperature threshold, a pre-judgment increment is added to the compensated spray command; when the rate of change is detected to be rapidly decreasing and the dimensionless temperature value is lower than the preset low temperature threshold, a pre-judgment increment is decremented from the compensated spray command; otherwise, the compensated spray command remains unchanged. The command obtained after the above operations is used as the adjusted spray command for subsequent integral anti-saturation processing and rate limiting processing.

[0069] Furthermore, the inverse function is obtained through on-site calibration and measurement: the actual flow rate is measured at three different opening points of the spray valve, and the correspondence between these three opening points and the flow rate is used to obtain three fitting coefficients through quadratic curve fitting. These three fitting coefficients constitute the parameters of the inverse function. In trend judgment, rapid rise means that the rate of change of dimensionless temperature value in three consecutive sampling periods exceeds the preset rise rate threshold, and rapid fall means that the rate of change of dimensionless temperature value in three consecutive sampling periods is lower than the preset fall rate threshold. The magnitude of the predicted increment is determined proportionally according to the absolute value of the rate of change. The larger the absolute value of the rate of change, the larger the predicted increment. The predicted increment increases or decreases in a fixed step size until it returns to zero after the trend is eliminated.

[0070] It should be noted that step 203 yielded a superimposed spray command of 0.620. Moving to step 204, the system executes inverse function compensation. The inverse function parameters determined during the initial calibration are: quadratic coefficient 0.20, linear coefficient 0.85, and constant term 0.03 (representing a 3% mechanical dead zone). The system substitutes the command 0.620 into the calculation formula: The process can be broken down as follows: the square term is 0.07688, and the linear term is 0.527. After adding the constant term, the exact result is 0.63388. The control system engineering truncation and rounding to three decimal places gives 0.634. Therefore, the compensated sprinkler command is 0.634.

[0071] The system then proceeds to trend prediction. The set high-temperature warning threshold is 0.550, and the rate of increase threshold is +0.015. The current temperature of 0.560 is already greater than 0.550. Simultaneously, the system retrieves the temperature change rates for the most recent three consecutive periods, which are +0.016, +0.018, and +0.020, all greater than the rate threshold of 0.015, indicating that the quench tower is undergoing accelerated thermal shock.

[0072] The system triggers a predictive intervention mechanism, calculating the increment using the following logarithmic predictive formula:

[0073] in, For dimensionless prediction of increments; current temperature change rate It is 0.020; operating condition proportional coefficient. Set to 0.020; rate amplification factor Set as (Engineering approximation value 85.914), ensuring the measuring point falls at the natural constant node; the continuous trigger count value m is 1 this time; safety step value. Set to 0.010.

[0074] Substitute into the calculation: rate amplification term The logarithmic part is e-1. The result is 1. The first part of the formula, the dynamic response value, is 0.020 multiplied by 1, which equals 0.020; the second part, the step increment, is 1 multiplied by 0.010, which equals 0.010. Adding the two together, we get the predicted increment for the current cycle as 0.030. The system adds the compensated command 0.634 to the predicted increment 0.030, outputting an adjusted spray command of 0.664, which is then sent to step 205.

[0075] 205. Perform integral anti-saturation processing and rate limiting processing on the adjusted spray command to obtain the spray command output to the actuator; at the same time, based on the sliding window statistical method, adaptively update the compensation parameters of the inverse function and the magnitude of the predicted increment. The statistical method uses the running data of multiple consecutive cycles to calculate the median or mean, without relying on the machine learning model.

[0076] Specifically, the adjusted spray command undergoes integral anti-saturation processing: it is determined whether the adjusted spray command exceeds the preset upper and lower limits. If it exceeds the upper limit, only the proportional derivative part is retained; if it exceeds the lower limit, only the proportional derivative part is retained as well. Integral accumulation is resumed when the command returns to the range, resulting in the anti-saturation processed spray command. The anti-saturation processed spray command then undergoes rate limiting processing: it is compared with the spray command of the previous cycle. If the difference between the two exceeds the preset maximum allowable rate of change, the current command is limited to the spray command of the previous cycle plus or minus this maximum rate of change. The spraying command is obtained by determining the allowable spraying rate and outputting it to the spraying actuator. The compensation parameters of the inverse function and the magnitude of the predicted increment are adaptively updated based on a sliding window statistical method: temperature overshoot data and valve feedback data from multiple consecutive operating cycles are collected. The median or mean of the temperature overshoot and the median or mean of the deviation between valve feedback and the command are calculated respectively. The magnitude of the predicted increment is adjusted based on the statistical results of the temperature overshoot, and the compensation parameters of the inverse function are adjusted based on the statistical results of the deviation between valve feedback and the command. The adjusted compensation parameters and magnitudes are used for the corresponding processing in the next operating cycle.

[0077] It should be noted that the system performs integral anti-saturation processing. The current command 0.664 is within the safe range of 0 to 1, and no over-limit saturation has occurred, so the integrator is working normally. Rate limiting processing is then performed; the specification limits the maximum allowable change in valve opening per cycle to 0.030. The system retrieves the command 0.620 issued in the previous cycle. The difference between the current command 0.664 and 0.620 is positive 0.044. Since the safety limit of 0.030 has been exceeded, the system triggers cutoff protection, limiting the maximum allowable increase in this cycle to 0.030. The system adds 0.030 to 0.620, resulting in a spray command output of 0.650, which drives the field pneumatic control valve.

[0078] The background adaptive program is enabled, and a sliding window is set to collect stable tracking data from the most recent 5 running cycles, as shown in Table 2 below: Table 2

[0079] Extracting the steady-state execution deviation data from the table for five cycles, and taking the median of 0.015 from the middle position in the sorted data, we can conclude that the valve exhibits a recent fixed mechanical hysteresis of 0.015.

[0080] The system automatically increases the compensation parameter of the constant term in the inverse function by 0.015. Simultaneously, the arithmetic mean of the temperature overshoot is calculated to be 0.0308. Considering that the normal allowable overshoot is only 0.015, a continuously high mean indicates that the conventional intervention is weak. Therefore, the system automatically increases the safety step threshold value for the predicted increment in step 204 from 0.010 to 0.012 to enhance the protection in the next cycle.

[0081] 206. When the system starts up or when the boiler load change exceeds the preset threshold, an initialization calibration process is performed. This process includes: forcing the spray valve to the preset reference opening, waiting for the flue gas temperature to reach a stable state and recording the stable temperature value; calculating and updating the zero-point offset of the inverse function based on the difference between the stable temperature value and the design reference temperature value, thereby obtaining the calibrated inverse function parameters; the calibrated inverse function parameters are used for subsequent dead zone and nonlinearity compensation.

[0082] Furthermore, the preset reference opening value is taken as the middle position of the spray valve's full stroke; the condition for waiting for the flue gas temperature to reach a stable state is: the rate of change of the dimensionless temperature value is continuously lower than the preset stability threshold for multiple consecutive sampling cycles; the zero-point offset is calculated by multiplying the difference between the stable temperature value and the design reference temperature value by a preset correction coefficient to obtain the offset value; the calibrated inverse function parameter is obtained by superimposing the offset value onto the constant term of the original inverse function and replacing the original inverse function parameter for subsequent operating cycles.

[0083] It should be noted that the boiler switched to hazardous waste treatment batches according to the production plan. Under the premise of ensuring safe operation, the overall thermal load was steadily increased by 20% through the feeding system. After the combustion conditions initially leveled off under the new load and the fluctuations in flue gas volume and temperature narrowed, the control system detected that the change in overall load exceeded the 15% trigger threshold. At this point, the system safely initiated the initialization calibration process, automatically suspending the routine closed-loop calculations in steps 201 to 205, and issued a forced command to fix the opening of the regulating valve at 50% (dimensionless command 0.500).

[0084] The system begins monitoring the rate of change of dimensionless temperature, with the criterion being that the rate of change remains below 0.005 for four consecutive cycles. The monitoring data is shown in Table 3 below: Table 3

[0085] From cycle 2 to cycle 5, the dimensionless temperature change rates were 0.004, 0.002, 0.001, and 0.001, respectively, all meeting the criterion of being below 0.005. At the end of cycle 5, the system determined that a new thermal equilibrium had been reached, and recorded a stable temperature value of 0.556.

[0086] According to the design manual, the reference dimensionless temperature should be 0.500 when the load is normal and the valve is 50% open. The measured value of 0.556 minus 0.500 yields a difference of positive 0.056, indicating an overall temperature rise under the new operating condition. The system calls the preset opening correction coefficient of 0.250, multiplies the temperature difference of 0.056 by 0.250, and calculates the zero-point offset as 0.014. The system retrieves the original inverse function constant term of 0.030 from step 204, adds 0.014 to the original constant term to obtain a new constant term of 0.044. The system replaces the original parameter with 0.044. After step 206 is completed, the system unlocks the valve and resumes normal control with the new parameters.

[0087] Figure 4 This is a schematic diagram of the structure of a boiler spray intelligent control device based on flue gas temperature closed-loop feedback provided in an embodiment of the present invention. The boiler spray intelligent control device 300 based on flue gas temperature closed-loop feedback may include a processor 301 and a memory 302. The memory 302 is used to store program instructions and / or data, and the processor 301 is used to execute the program instructions stored in the memory 302, thereby implementing the method in the above-described method embodiment.

[0088] Optionally, the memory 302 and the processor 301 are coupled. The coupling is an indirect coupling or communication connection between devices, units, or modules, and can be electrical, mechanical, or other forms, for information interaction between devices, units, or modules.

[0089] Optionally, the boiler spray intelligent control device 300 based on closed-loop feedback of flue gas temperature may further include a communication interface 303. The communication interface 303 is used to communicate with other devices through a transmission medium, for example, transmitting signals received from other communication devices to the processor 301, or transmitting signals from the processor 301 to other communication devices. The communication interface 303 may be a transceiver or an interface circuit, such as a transceiver circuit or a transceiver chip.

[0090] This application embodiment does not limit the specific connection medium between the processor 301, memory 302, and communication interface 303. This application embodiment... Figure 4 The processor 301, memory 302, and communication interface 303 are connected via a bus 304. Figure 4 The connections between other components are shown in bold and are for illustrative purposes only, not as limiting information. The bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0091] The present invention also provides a boiler spray intelligent control device based on flue gas temperature closed-loop feedback. The boiler spray intelligent control device based on flue gas temperature closed-loop feedback includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the boiler spray intelligent control method based on flue gas temperature closed-loop feedback in the above embodiments.

[0092] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the intelligent control method for boiler spraying based on closed-loop feedback of flue gas temperature.

[0093] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0094] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0095] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent control of boiler spraying based on closed-loop feedback of flue gas temperature, characterized in that, include: The collected flue gas temperature and spray valve position were subjected to dimensionless normalization to obtain dimensionless temperature value and dimensionless valve position value. The adaptive control parameters are determined based on the dimensionless temperature value and its rate of change, and the closed-loop control calculation is performed based on the temperature deviation between the dimensionless temperature value and the temperature setpoint to obtain the initial spray command. The feedforward compensation amount is calculated based on the deviation between the oxygen content of the flue gas and the set oxygen content. The feedforward compensation amount is then combined with the initial spray command to obtain the superimposed spray command. The superimposed spray command is nonlinearly compensated using a pre-established inverse function to obtain the compensated spray command; then, the trend is predicted based on the rate of change of the dimensionless temperature value, and the predicted increment is superimposed on the compensated spray command to obtain the adjusted spray command. The adjusted spray command is restricted and then output to the spray actuator; at the same time, the continuous operation data is statistically analyzed based on statistical methods, and the compensation parameters of the inverse function and the amplitude of the predicted increment are adaptively updated.

2. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 1, characterized in that, include: The difference between the measured value of the oxygen content in the flue gas and the set value of the oxygen content is calculated as the oxygen content deviation; The oxygen content deviation is multiplied by a feedforward coefficient to obtain the feedforward compensation amount, wherein the feedforward coefficient is adaptively adjusted by statistically analyzing the sign relationship between the temperature deviation and the oxygen content deviation online. The feedforward compensation amount is added to the initial spray command to obtain the superimposed spray command.

3. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 2, characterized in that, The adaptive adjustment process of the feedforward coefficient includes: The oxygen content deviation and the temperature deviation are collected over multiple consecutive control cycles. If the oxygen content deviation and the temperature deviation have the same sign, then the current feedforward coefficient is increased by a preset step increment. If the oxygen content deviation has the opposite sign to the temperature deviation, then the current feedforward coefficient is reduced by a preset step decrease. Wherein, both the step increment and the step decrement are fixed values.

4. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 1, characterized in that, include: The superimposed spray command is input into the inverse function, and the valve dead zone is eliminated and the gain nonlinearity is corrected through the inverse function to obtain the compensated spray command. When the rate of change of the dimensionless temperature value is detected to be rising rapidly and the dimensionless temperature value exceeds the preset high temperature threshold, the predicted increment is added to the compensated spray command. When it is detected that the rate of change of the dimensionless temperature value is in a state of rapid decline and the dimensionless temperature value is lower than the preset low temperature threshold, the predicted increment is reduced on the compensated spray command. Otherwise, the compensated spray command remains unchanged; The above-mentioned adjusted results shall serve as the adjusted spraying instructions.

5. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 4, characterized in that, The inverse function is obtained through on-site calibration: the actual flow rate is measured at three different opening points of the spray actuator, and the correspondence between the three opening points and the flow rate is used to obtain three fitting coefficients through quadratic curve fitting. The fitting coefficients constitute the compensation parameters of the inverse function. The rapid rise refers to the rate of change of the dimensionless temperature value exceeding a preset rise rate threshold for three consecutive sampling cycles, and the rapid fall refers to the rate of change of the dimensionless temperature value being lower than a preset fall rate threshold for three consecutive sampling cycles. The magnitude of the predicted increment is positively correlated with the absolute value of the rate of change of the dimensionless temperature value, and the predicted increment increases or decreases with a fixed step size until the trend is eliminated and returns to zero.

6. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 5, characterized in that, The predicted increment is calculated using the following logarithmic formula: ; in, To predict the increment; This represents the rate of change of the dimensionless temperature value during the current sampling period. This is a preset operating condition ratio coefficient; is the preset rate amplification factor; m is the count value for continuous trigger trend judgment; This is the step value with a fixed step size.

7. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 1, characterized in that, include: The adjusted spray command is subjected to integral anti-saturation processing: it is determined whether the adjusted spray command exceeds the preset upper or lower limit. If it exceeds the upper or lower limit, only the proportional and derivative parts in the closed-loop control calculation are retained, and integral accumulation is restored when the command returns to the range between the upper and lower limits to obtain the anti-saturation processed spray command. The spray command after the anti-saturation treatment is rate-limited: the spray command after the anti-saturation treatment is compared with the spray command of the previous control cycle. If the difference between the two exceeds the preset maximum allowable change rate, the current output is limited to the spray command of the previous control cycle plus or minus the maximum allowable change rate, so as to obtain the spray command output to the spray actuator.

8. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 1, characterized in that, include: Using a sliding window statistical method, temperature overshoot and valve command deviation are collected for multiple consecutive operating cycles. The median or mean of the temperature overshoot and the median or mean of the valve command deviation are calculated respectively. The amplitude of the predicted increment is adaptively adjusted based on the statistical results of the temperature overshoot. The compensation parameters of the inverse function are adaptively adjusted based on the statistical results of the valve command deviation. The updated compensation parameters and amplitudes are used for the adjustment processing in the next operating cycle.

9. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 1, characterized in that, It also includes the initial calibration process: When the system starts up or when the boiler load change exceeds a preset threshold, the spray valve is forced to a preset reference opening, and the stable temperature value under the stable state is recorded after the flue gas temperature reaches a stable state. Based on the difference between the stable temperature value and the design reference temperature value, the zero-point offset of the inverse function is calculated and updated to obtain the calibrated inverse function compensation parameters, which are used for subsequent nonlinear compensation.

10. The intelligent boiler spray control method based on closed-loop feedback of flue gas temperature according to claim 9, characterized in that, The reference opening is the midpoint of the spray valve's full stroke; The condition for determining the stable state is that the rate of change of the dimensionless temperature value is continuously lower than a preset stability threshold within multiple consecutive sampling periods. The zero-point offset is calculated as follows: the difference between the stable temperature value and the design reference temperature value is multiplied by a preset correction coefficient to obtain the offset value; the offset value is then added to the constant term of the inverse function to replace the original compensation parameter.

Citation Information

Patent Citations

  • Method and device for optimizing smoke exhaust temperature of power plant boiler

    CN107145981A

  • Intelligent boiler control method

    CN111396846A