SCR denitration system ammonia injection control method based on big data prediction model

Abstract

The invention relates to the technical field of digital processing, in particular to an SCR (Selective Catalytic Reduction) denitration system ammonia spraying control method based on a big data prediction model, which comprises the following steps: in order to optimize a denitration system, collecting nitrogen oxide concentration, ammonia spraying flow, power and flue gas flow velocity data of an inlet and an outlet of an SCR reactor, and obtaining the maximum ammonia flow. By analyzing these data, a first time lag between the power and the nitrogen oxide concentration, a third time lag between the inlet and outlet concentrations, and a fourth time lag between the ammonia injection flow and the outlet concentration are determined. And calculating a second time lag value according to a difference value between the fourth time lag value and the third time lag value, and adjusting the ammonia spraying flow according to the second time lag value. According to the invention, the ammonia spraying flow can be effectively regulated and controlled, and the denitration efficiency and the system operation stability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an ammonia injection control method for an SCR denitration system based on a big data prediction model. Background Art

[0002] With the increasing global attention to environmental protection and air quality, nitrogen oxides (nitrogen oxide gas) as one of the main air pollutants, its emission control has become an important goal of various enterprises. However, the traditional SCR system faces many challenges in operation, especially in the optimization of ammonia injection and the stable control of nitrogen oxide gas concentration. In the actual denitrification process, there is a time delay between the injection of ammonia and the concentration of nitrogen oxide gas. This lag makes it difficult for the traditional control method to adjust the ammonia injection amount in real time, resulting in the outlet nitrogen oxide gas concentration cannot effectively follow the set value; and in order to ensure the complete reduction of nitrogen oxide gas, the traditional SCR system often adopts the method of excessive ammonia injection, which not only leads to the waste of ammonia water, but also may cause ammonia escape, causing secondary pollution to the environment. At present, the SCR system does not make full use of historical operation data in the control process, lacks a prediction model based on big data analysis to adjust the ammonia injection amount in advance, and the system operation efficiency is low. Therefore, through the analysis of historical data and real-time monitoring, the optimization control of ammonia injection amount can be realized, thereby effectively reducing nitrogen oxide gas emissions and reducing the use of ammonia, which is a problem that the current SCR technology needs to solve. Summary of the invention

[0003] The present invention provides an ammonia injection control method for an SCR denitration system based on a big data prediction model to solve the existing problems.

[0004] The SCR denitration system ammonia injection control method based on the big data prediction model of the present invention adopts the following technical solutions: An embodiment of the present invention provides an ammonia injection control method for an SCR denitration system based on a big data prediction model, the method comprising the following steps: During the period from the beginning of the reaction process to the current moment, obtain the time series data of the nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor, obtain the ammonia injection flow, power, and flue gas flow rate of the SCR denitrification system; and obtain the maximum ammonia flow when the valve is fully opened; Acquire a first time lag value according to the power data sequence and the nitrogen oxide gas concentration data sequence; According to the nitrogen oxide gas concentration time series data and the flue gas flow rate at the inlet and outlet of the SCR reactor, a third time lag value is obtained; according to the ammonia injection flow rate, the nitrogen oxide gas concentration time series data and the flue gas flow rate at the outlet of the SCR reactor, a fourth time lag value is obtained; according to the fourth time lag value and the third time lag value, a second time lag value is obtained; The ammonia injection flow rate is regulated according to the size of the second time lag value and the maximum ammonia flow rate when the valve is fully opened.

[0005] Preferably, the obtaining of the first time lag value according to the power data sequence and the nitrogen oxide gas concentration data sequence comprises the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The Pearson correlation coefficient of the time series data of all power and nitrogen oxide gas concentration at the inlet of all SCR reactors in the target period at the moment is taken as the The correlation coefficient at each moment; Obtain the correlation coefficient at each moment, and record the largest correlation coefficient among all moments as the first time lag value of the flue gas from the boiler to the SCR reactor .

[0006] Preferably, the step of obtaining the third time lag value according to the time series data of the nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor and the flue gas flow rate comprises the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The standard deviation of all ammonia injection flow rates within the target period at time ; The time series data of nitrogen oxide gas concentration at the inlet of SCR reactor Within the target period of time, , as the first The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at the moment, where , as well as They are the first and second time series data of nitrogen oxide gas concentration at the inlet of SCR reactor. Within the target period of time, moment, The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at each moment; In the Within the target period of time, The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment is The ratio of The hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment; Get the The time corresponding to the local maximum value of the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at all times within the target period of time is taken as the time of the first A reference time within the target period of a moment; Get the For each target time period at a reference time within a target time period at each reference time, calculate the Pearson correlation coefficient between the nitrogen oxide gas concentration time series data at the inlet and outlet of all SCR reactors within the target time period at each reference time as the second correlation coefficient at each reference time, and obtain the largest of the second correlation coefficients at all reference times as the third maximum significant time lag value; The reference time corresponding to the maximum second correlation coefficient is recorded as The target moment corresponding to each moment; In the In the target period of each moment, the first target time is obtained according to the number of reference moments, the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment, and the flue gas flow rate. The third time lag value corresponding to the moment.

[0007] Preferably, the In the target period of each moment, the first target time is obtained according to the number of reference moments, the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment, and the flue gas flow rate. The third time lag value corresponding to the moment includes the following specific steps: Get the The sum of the normalized values ​​of the lag significance of all reference moments within the target period of the moment; Get the third largest significant time lag value and the The ratio of the flue gas flow rate value at the target time corresponding to the first time is further obtained; the ratio is The product of the flue gas velocity values ​​at each moment; The first The sum of the normalized values ​​of the hysteresis significance of all reference moments in the target period of the moment, and the ratio of the ratio to the first The product of the flue gas velocity values ​​at the time is recorded as The third time lag value corresponding to the moment.

[0008] Preferably, the step of obtaining the fourth time lag value according to the ammonia injection flow rate, the time series data of the nitrogen oxide gas concentration at the outlet of the SCR reactor and the flue gas flow rate comprises the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The standard deviation of the time series data of nitrogen oxide gas concentration at the outlet of all SCR reactors within the target period at the moment ; At the injection flow rate Within the target period of time, , as the first The second-order difference of the ammonia injection flow rate at each moment, where: , as well as They are the ammonia injection flow rate Within the target period of time, moment, The moment and Ammonia injection flow value at a certain moment; In the Within the target period of time, The second-order difference of the ammonia injection flow rate at each moment is The ratio of The hysteresis significance of the ammonia injection flow rate at each moment; Get the The time corresponding to the local maximum value of the hysteresis significance of the ammonia injection flow rate at all times within the target period of a moment is taken as the reference time; Get the For each target period at a reference time within a target period at each reference time, calculate the Pearson correlation coefficient between the nitrogen oxide gas concentration time series data at all SCR outlets and the ammonia injection flow rate within the target period at each reference time as the third correlation coefficient at each reference time, and obtain the largest of the third correlation coefficients at all reference times as the fourth maximum significant time lag value; The reference time corresponding to the largest third correlation coefficient is recorded as The target moment corresponding to each moment; In the In the target period of each moment, the first The fourth time lag value corresponding to the moment.

[0009] Preferably, the In the target period of each moment, the first The fourth time lag value corresponding to the moment includes the following specific steps: Get the target time period of all times The sum of the normalized values ​​of the lag significance at the reference moments; Get the fourth largest significant time lag value and the The ratio of the flue gas flow rate value at the target time corresponding to the first time is further obtained; the ratio is The product of the flue gas velocity values ​​at each moment; The first The sum of the normalized values ​​of the hysteresis significance of all reference moments in the target period of the moment, and the ratio of the ratio to the first The product of the flue gas velocity values ​​at the time is recorded as The fourth time lag value at this moment.

[0010] Preferably, the step of obtaining the second time lag value according to the fourth time lag value and the third time lag value comprises the following specific steps: The second time lag value is The fourth time lag value corresponding to the moment minus the The difference of the third time lag values ​​corresponding to the moments.

[0011] Preferably, the step of regulating the ammonia injection flow rate according to the second time lag value and the maximum ammonia flow rate when the valve is fully opened comprises the following specific steps: The second time lag is ; When the second time lag value is less than or equal to zero, use The nitrogen oxide gas concentration at the inlet of the SCR reactor at the time The nitrogen oxide gas concentration at the inlet of the SCR reactor at the moment is used to regulate the ammonia injection flow rate; When the second time lag value is greater than zero, the predicted gas value at the reactor inlet is obtained according to the power, and the valve opening percentage is obtained according to the predicted gas value at the reactor inlet and the maximum ammonia flow when the valve is fully opened.

[0012] Preferably, the step of obtaining the predicted gas value at the reactor inlet according to the power comprises the following specific steps: In the time series data of nitrogen oxide gas concentration at the inlet of the SCR reactor, As the The first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor at the moment, where and is the time series data of nitrogen oxide gas concentration at the inlet of SCR reactor The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at a certain moment; In power, As the The first-order difference of power at each moment is and The power The moment and The power value at a moment; Get At the moment, the first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor is The ratio of the first-order difference of power at the time; further obtaining the mean of the ratio at all times, recorded as the influence coefficient; Build from 0 to The second time lag value sequence of is any value in the second time lag value sequence; The first time lag is ; Get the influence coefficient and At the moment, the product of the first-order difference of power is further obtained, and the product and At the time, the sum of the nitrogen oxide gas concentration values ​​at the inlet of the SCR reactor is recorded as the predicted sum value; Get the sum of all predicted values ​​for all values ​​in the second time lag value sequence, recorded as The predicted value of gas at the reactor inlet at the time .

[0013] Preferably, the method of obtaining the valve opening percentage according to the predicted gas value at the reactor inlet and the maximum ammonia flow rate when the valve is fully opened comprises the following specific steps: According to the nitrogen oxide gas concentration data under the stable output state of the historical nitrogen oxide gas concentration, the corresponding optimal ammonia flow rate is obtained; Get pending parameters , The predicted value of gas at the reactor inlet at the time and the flue gas velocity Flue gas velocity value at the moment The product of the three; further according to the product of the three and the flow bias term The sum of the values, constructing and corresponding to the moment The equivalent relationship of the optimal ammonia flow rate under Based on the equivalent relationship and the optimal ammonia flow rate at all historical moments, the gas prediction value at the reactor inlet and the flue gas velocity Flue gas velocity value at the moment , perform function fitting and obtain and The value of is used to build a regression model; According to the regression model, the gas prediction value and flue gas flow rate value at the current moment are input to obtain the required ammonia flow at the current moment in real time; The percentage of the required ammonia flow rate at the current moment to the maximum ammonia flow rate when the valve is fully opened is recorded as the valve opening percentage at the current moment.

[0014] The beneficial effect of the technical solution of the present invention is that the system can more accurately predict the nitrogen oxide gas concentration at the reactor inlet by comprehensively considering the time lag value of the flue gas from the boiler to the SCR reactor, the time lag value of the change of the nitrogen oxide gas concentration between the SCR inlet and outlet, and the time lag value of the ammonia injection flow data and the change of the outlet nitrogen oxide gas concentration. Accurate time lag compensation and ammonia flow prediction are achieved, and the real-time control capability of the ammonia injection valve is significantly improved. This method not only reduces the excessive use of ammonia, reduces operating costs and environmental risks, but also enhances the stability and adaptability of the system, and provides an effective solution for the SCR denitrification system in dealing with boiler load fluctuations and nitrogen oxide gas emission control, and promotes technological progress and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 The present invention is a flowchart of the steps of the ammonia injection control method of the SCR denitrification system based on the big data prediction model. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, characteristics and effects of the SCR denitration system ammonia injection control method based on the big data prediction model proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, 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 belongs.

[0019] The specific scheme of the ammonia injection control method of the SCR denitrification system based on the big data prediction model provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a flow chart of the steps of an SCR denitration system ammonia injection control method based on a big data prediction model provided by an embodiment of the present invention, the method comprising the following steps: Step S001: During the period from the start of the reaction process to the current moment, obtain the time series data of the nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor, obtain the ammonia injection flow, power, and flue gas flow rate of the SCR denitrification system; and obtain the maximum ammonia flow when the valve is fully opened.

[0021] The nitrogen oxide gas concentration time series data at the inlet and outlet of the SCR reactor are obtained; wherein the nitrogen oxide gas concentration value is obtained in real time by optical nitrogen oxide gas concentration sensors at the inlet and outlet of the SCR reactor.

[0022] The ammonia injection flow rate is obtained; wherein the ammonia injection flow rate value is obtained in real time through a flow sensor installed at the ammonia injection valve.

[0023] Obtain power; wherein the power value is obtained in real time by monitoring the boiler in the SCR reactor through a power sensor.

[0024] Obtain the flue gas flow rate; wherein the flue gas flow rate value is obtained in real time by installing a flow meter in the flue.

[0025] Get the maximum ammonia flow rate with the valve fully open.

[0026] It should be noted that the above time series data are all equal-length sequences, the time series data collection duration is the time series data from the beginning of the reaction process to the current moment, and the time series data collection frequency is 30 seconds. Data at each moment can be obtained through data fitting.

[0027] It should be noted that the full name of the SCR denitrification system reactor is the selective catalytic reduction reactor. The ammonia injection system in the reactor is a core part of the SCR denitrification technology. It is responsible for spraying the reducing agent (such as ammonia) evenly into the flue gas so as to react with the nitrogen oxide gas in the flue gas to reduce the emission of nitrogen oxide gas.

[0028] Step S002: Obtain a first time lag value according to the power data sequence and the nitrogen oxide gas concentration data sequence.

[0029] In order to predict the change of nitrogen oxide gas concentration during the reaction of the SCR reactor, it is necessary to consider the first time lag value of the change of boiler load compared to the inlet of the SCR reactor.

[0030] In the period from the beginning of the reaction process to the current moment, from the Start to go back in time The time period of time is recorded as It should be noted that for the target period starting from the reaction starting point and extending backward The partial time period between the moments is not considered as the analysis object in this example. It is a preset value, which is 50 in this embodiment.

[0031] Calculate the The Pearson correlation coefficient of the time series data of all power and nitrogen oxide gas concentration at the inlet of all SCR reactors in the target period at the moment is taken as the Further, the correlation coefficient of each moment is obtained, and the largest correlation coefficient among all moments is recorded as the first time lag value of the flue gas from the boiler to the SCR reactor. The acquisition of the Pearson correlation coefficient is a well-known technique and will not be described in detail here.

[0032] It should be noted that the time lag value refers to the time required from the boiler load change to the generation of nitrogen oxide gas, and then to the change appearing at the inlet of the SCR reactor. Understanding the lag is crucial to optimizing the ammonia injection control of the SCR system, because it can more accurately predict the change in the concentration of nitrogen oxide gas and adjust the amount of ammonia injection accordingly to ensure that the denitrification efficiency is maximized while reducing ammonia slip, thereby effectively controlling pollutant emissions and meeting environmental protection requirements. For example, when the boiler load needs to be increased, the system will increase the supply of fuel and air accordingly. As the combustion temperature and pressure increase, the generation rate of nitrogen oxide gas will usually increase; on the contrary, when the boiler load needs to be reduced, the system will reduce the fuel supply and lower the combustion temperature, thereby reducing the generation of nitrogen oxide gas.

[0033] Step S003: Obtain a third time lag value according to the time series data of nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor and the flue gas flow rate; obtain a fourth time lag value according to the ammonia injection flow rate, the time series data of nitrogen oxide gas concentration at the outlet of the SCR reactor and the flue gas flow rate; obtain a second time lag value according to the fourth time lag value and the third time lag value.

[0034] In order to predict the change of nitrogen oxide gas concentration at the ammonia injection valve and the inlet in the reactor, it is necessary to further consider the second time lag between the change of nitrogen oxide gas concentration and the change of ammonia injection flow rate. , the second time lag value, through the third time lag value and the fourth lag Acquire. According to the time series data of nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor, the third time lag value is acquired. According to the time series data of the ammonia injection flow rate and the nitrogen oxide gas concentration at the outlet of the SCR reactor, the fourth time lag value is acquired.

[0035] Among them, the second time lag value is the difference between the fourth time lag value and the third time lag value. It should be noted that since ammonia requires time to flow and mix with nitrogen oxide gas, the time required in this process cannot be directly obtained. Therefore, this embodiment obtains the estimated time from ammonia injection to participation in the reaction by the difference between the time lag value between the nitrogen oxide gas concentration at the inlet and the ammonia injection flow rate data and the nitrogen oxide gas concentration data at the outlet.

[0036] In this embodiment, a process for determining the third time lag value may be: Calculate the The standard deviation of all ammonia injection flow rates within the target period at time ; The time series data of nitrogen oxide gas concentration at the inlet of SCR reactor Within the target period of time, , as the first The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at the moment, where , as well as They are the first and second time series data of nitrogen oxide gas concentration at the inlet of SCR reactor. Within the target period of time, moment, The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at the time. In this implementation, the first and second moments in the target time period are not analyzed.

[0037] In the Within the target period of time, The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment is The ratio of The hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment.

[0038] Get the The time corresponding to the local maximum value of the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at all times within the target period of time is taken as the time of the first The acquisition of the local maximum value is a well-known technique and will not be described in detail here.

[0039] Get the For each target time period at the reference time within the target time period at the reference time, calculate the Pearson correlation coefficient between the nitrogen oxide gas concentration time series data at the inlet and outlet of all SCR reactors within the target time period at each reference time as the second correlation coefficient at each reference time, obtain the largest of the second correlation coefficients at all reference times as the third maximum significant time lag value; record the reference time corresponding to the largest second correlation coefficient as the The target moment corresponding to the moment.

[0040] In the In the target period of each moment, the first target time is obtained according to the number of reference moments, the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment, and the flue gas flow rate. The third time lag value corresponding to the moment.

[0041] This embodiment provides a method for determining the The formula for the third time lag value corresponding to the moment can be:

[0042] in, For the The third time lag value corresponding to the moment; For the The number of reference moments in the target period of moments, For the The target period of time The lag significance of the reference time, is the third most significant time lag value, For the The flue gas velocity value at the target time corresponding to the moment is: For the The flue gas velocity value at the moment, is a linear normalization function used to normalize data values ​​to between 0 and 1. To round up.

[0043] It should be noted that since the change in the concentration of nitrogen oxide gas after passing through the SCR reactor will vary due to different ammonia concentrations, its time lag does not depend solely on a single factor. Based on this, it may not be reliable to calculate the time lag using only the correlation coefficient. In order to reduce the impact of different ammonia concentrations on the time lag value, this embodiment selects a stable ammonia injection time period through the above steps to more accurately estimate the time lag of the change in the nitrogen oxide gas concentration at the inlet and outlet.

[0044] In this embodiment, a process for determining the fourth time lag value is provided as follows: Calculate the The standard deviation of the time series data of nitrogen oxide gas concentration at the outlet of all SCR reactors within the target period at the moment ; At the injection flow rate Within the target period of time, , as the first The second-order difference of the ammonia injection flow rate at each moment, where: , They are the ammonia injection flow rate Within the target period of time, moment, The moment and The ammonia injection flow value at each moment. In this implementation, the first and second moments in the target period are not analyzed.

[0045] In the Within the target period of time, The second-order difference of the ammonia injection flow rate at each moment is The ratio of The degree of lag of ammonia injection flow at each moment.

[0046] Get the The time corresponding to the local maximum value in the hysteresis significance degree of the ammonia injection flow rate at all times within the target period of the moment is taken as the reference time.

[0047] Get the For each target period at a reference time within a target period at a certain moment, calculate the Pearson correlation coefficient between the time series data of the nitrogen oxide gas concentration at all SCR outlets and the ammonia injection flow rate within the target period at each reference time as the third correlation coefficient at each reference time, obtain the largest of the third correlation coefficients at all reference times as the fourth maximum significant time lag value; record the reference time corresponding to the largest third correlation coefficient as the The target moment corresponding to the moment.

[0048] In the In the target period of each moment, the first This embodiment provides a method for determining the fourth time lag value corresponding to the first moment. The formula for the fourth time lag value corresponding to the moment can be:

[0049] in, For the The fourth time lag value corresponding to the moment; For the The number of reference moments within the target period of the moment, For the The target period of time The lag significance of the reference time is is the fourth largest significant time lag value, For the The flue gas velocity value at the target time corresponding to the moment is: For the The flue gas velocity value at the moment, is a linear normalization function used to normalize data values ​​to between 0 and 1. To round up.

[0050] It should be noted that: the above steps take into account that the change in nitrogen oxide gas concentration at the outlet of the SCR reactor not only depends on the nitrogen oxide gas concentration at the inlet, but also that based on different ammonia injection flow rates, the change in nitrogen oxide gas concentration at the outlet will also decrease with the increase of ammonia injection flow rate, so as to obtain the fourth time lag value.

[0051] Step S004: regulating the ammonia injection flow rate according to the second time lag value and the maximum ammonia flow rate when the valve is fully opened.

[0052] The ammonia injection flow rate is regulated according to the size of the second time lag value.

[0053] Specifically, when the second time lag value is less than or equal to zero , the ammonia injection flow rate can be controlled after the flue gas reaches the reactor inlet. Value adjustment at the moment Ammonia injection flow at the time.

[0054] It should be noted that changes in boiler load will directly affect the combustion conditions in the furnace, thereby changing the amount of nitrogen oxides generated. When operating at high load, in order to maintain the nitrogen oxide emission concentration up to standard, the ammonia injection flow rate needs to be increased; while at low load, the ammonia injection flow rate can be appropriately reduced. This control method uses real-time monitoring data for feedback control to ensure that the removal efficiency of nitrogen oxides is maximized, while avoiding secondary pollution caused by excessive use of ammonia. The nitrogen oxide gas concentration at the inlet of the SCR reactor at the time It is a known operation to adjust the ammonia injection flow rate according to the nitrogen oxide gas concentration at the inlet of the SCR reactor at the time.

[0055] When the second time lag value is greater than zero , indicating that there is a significant time delay between the change in NOx gas concentration at the inlet of the SCR reactor and the change in the ammonia injection flow rate. In this case, relying solely on the NOx gas concentration data at the current moment or the previous moment cannot accurately reflect the upcoming change in NOx gas concentration. Therefore, in order to effectively control the ammonia injection flow rate, it is necessary to predict the NOx gas concentration at the inlet of the SCR reactor at a future moment. This prediction can help adjust the ammonia injection flow rate in advance to better match the NOx gas concentration at a future moment, thereby improving the denitrification efficiency and reducing ammonia slip.

[0056] Considering that the nitrogen oxide gas concentration value for regulating the ammonia injection flow is not only related to the data changes at historical times, but also that the nitrogen oxide gas concentration in the output flue gas will change with the difference in the corresponding load during the combustion process in the boiler, the prediction is considered based on the changes in the boiler load.

[0057] Specifically, the influence coefficient is obtained according to the power; the gas prediction value at the reactor inlet is obtained according to the influence coefficient; a regression model is constructed according to the gas prediction value at the reactor inlet; and the valve opening percentage is obtained according to the regression model.

[0058] According to the power acquisition influence coefficient, this implementation provides a process for determining, which may be: In the time series data of nitrogen oxide gas concentration at the inlet of the SCR reactor, As the The first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor at the moment, where and is the time series data of nitrogen oxide gas concentration at the inlet of SCR reactor The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at the moment. In this embodiment, the first moment is not analyzed.

[0059] In power, As the The first-order difference of power at each moment, where and The power The moment and The power value at each moment. In this implementation, the first moment is not analyzed.

[0060] Get At the moment, the first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor is The ratio of the first-order difference of power at the time instant is obtained. The average of the ratio at all times is further obtained and recorded as the influence coefficient.

[0061] Among them, the predicted value of the gas at the reactor inlet is obtained according to the influence coefficient. This embodiment provides a process for determining, which can be: Build from 0 to The second time lag value series ,definition is any value in the second time lag value sequence.

[0062] Calculate the influence coefficient and The product of the first-order difference of power at the time is further obtained. At the time, the sum of the nitrogen oxide gas concentration values ​​at the inlet of the SCR reactor is calculated and recorded as the predicted sum value.

[0063] Get the sum of all predicted values ​​for all values ​​in the second time lag value sequence, recorded as The predicted value of gas at the reactor inlet at the time .

[0064] Among them, a regression model is constructed according to the predicted value of the gas at the reactor inlet. This embodiment provides a process for determination, which can be: According to the nitrogen oxide gas concentration data under the stable output state of the historical nitrogen oxide gas concentration, the corresponding optimal ammonia flow rate is obtained.

[0065] It should be noted that the optimal ammonia flow rate is the optimal ammonia flow rate at each moment corresponding to the nitrogen oxide gas concentration in each reaction under a stable output state in multiple cycles of the same reaction.

[0066] According to the optimal ammonia flow rate at all historical moments, and the gas prediction value and flue gas flow rate value at the reactor inlet at the corresponding moment, a regression model is constructed. This embodiment provides a method for constructing a regression model as follows:

[0067] In the formula, For the The optimal ammonia flow rate at the moment, is a parameter to be determined; for The predicted value of gas at the reactor inlet at the time; For the Flue gas velocity value at a certain moment; is the flow bias term.

[0068] According to all historical moments Data, function fitting is performed on it, and then the and to construct a regression model.

[0069] According to the regression model, the gas prediction value and flue gas flow rate value at the current moment are input to obtain the required ammonia flow at the current moment in real time.

[0070] Among them, according to the regression model, the valve opening percentage is obtained. This implementation provides a process for determining, which can be: The percentage of the required ammonia flow rate at the current moment to the maximum ammonia flow rate when the valve is fully opened is recorded as the valve opening percentage at the current moment.

[0071] In each control cycle, the system updates the valve opening percentage in real time based on the current NOx gas concentration and the calculated required ammonia flow rate. By using PID control or other advanced control algorithms, the valve opening at any time is adjusted to ensure that the target flow rate is achieved. For example, if the actual flow rate detected is lower than the set target flow rate, the control system will automatically increase the valve opening; conversely, if the actual flow rate is higher than the target flow rate, the control system will reduce the valve opening. This dynamic adjustment mechanism ensures precise control and optimized operation of the process.

[0072] So far, the present invention is completed.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for controlling ammonia injection in an SCR denitration system based on a big data prediction model, characterized in that: The method comprises the following steps: During the period from the beginning of the reaction process to the current moment, obtain the time series data of the nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor, obtain the ammonia injection flow, power, and flue gas flow rate of the SCR denitrification system; and obtain the maximum ammonia flow when the valve is fully opened; Acquire a first time lag value according to the power data sequence and the nitrogen oxide gas concentration data sequence; According to the nitrogen oxide gas concentration time series data and the flue gas flow rate at the inlet and outlet of the SCR reactor, a third time lag value is obtained; according to the ammonia injection flow rate, the nitrogen oxide gas concentration time series data and the flue gas flow rate at the outlet of the SCR reactor, a fourth time lag value is obtained; according to the fourth time lag value and the third time lag value, a second time lag value is obtained; The ammonia injection flow rate is regulated according to the size of the second time lag value and the maximum ammonia flow rate when the valve is fully opened.

2. The SCR denitration system ammonia injection control method based on the big data prediction model according to claim 1 is characterized in that: The step of obtaining the first time lag value according to the power data sequence and the nitrogen oxide gas concentration data sequence includes the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The Pearson correlation coefficient of the time series data of all power and nitrogen oxide gas concentration at the inlet of all SCR reactors in the target period at the moment is taken as the The correlation coefficient at each moment; Obtain the correlation coefficient at each moment, and record the largest correlation coefficient among all moments as the first time lag value of the flue gas from the boiler to the SCR reactor .

3. The SCR denitration system ammonia injection control method based on the big data prediction model according to claim 1 is characterized in that: The step of obtaining the third time lag value according to the time series data of the nitrogen oxide gas concentration at the inlet and outlet of the SCR reactor and the flue gas flow rate includes the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The standard deviation of all ammonia injection flow rates within the target period at time ; The time series data of nitrogen oxide gas concentration at the inlet of SCR reactor Within the target period of time, , as the first The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at the moment, where , as well as They are the first and second time series data of nitrogen oxide gas concentration at the inlet of SCR reactor. Within the target period of time, moment, The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at each moment; In the Within the target period of time, The second-order difference of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment is The ratio of The hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment; Get the The time corresponding to the local maximum value of the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at all times within the target period of time is taken as the time of the first A reference time within the target period of a moment; Get the For each target time period at a reference time within a target time period at each reference time, calculate the Pearson correlation coefficient between the nitrogen oxide gas concentration time series data at the inlet and outlet of all SCR reactors within the target time period at each reference time as the second correlation coefficient at each reference time, and obtain the largest of the second correlation coefficients at all reference times as the third maximum significant time lag value; The reference time corresponding to the maximum second correlation coefficient is recorded as The target moment corresponding to each moment; In the In the target period of each moment, the first target time is obtained according to the number of reference moments, the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment, and the flue gas flow rate. The third time lag value corresponding to the moment.

4. The ammonia injection control method for the SCR denitration system based on the big data prediction model according to claim 3 is characterized in that: As stated in In the target period of each moment, the first target time is obtained according to the number of reference moments, the hysteresis significance of the nitrogen oxide gas concentration at the inlet of the SCR reactor at each moment, and the flue gas flow rate. The third time lag value corresponding to the moment includes the following specific steps: Get the The sum of the normalized values ​​of the lag significance of all reference moments within the target period of the moment; Get the third largest significant time lag value and the The ratio of the flue gas flow rate value at the target time corresponding to the first time is further obtained; the ratio is The product of the flue gas velocity values ​​at each moment; The first The sum of the normalized values ​​of the hysteresis significance of all reference moments in the target period of the moment, and the ratio of the ratio to the first The product of the flue gas velocity values ​​at the time is recorded as The third time lag value corresponding to the moment.

5. The ammonia injection control method for the SCR denitration system based on the big data prediction model according to claim 1 is characterized in that: The step of obtaining the fourth time lag value according to the ammonia injection flow rate, the nitrogen oxide gas concentration time series data at the outlet of the SCR reactor and the flue gas flow rate includes the following specific steps: In the period from the beginning of the reaction process to the current moment, from the Start to trace back to the preset value at the moment The time period of time is recorded as target time period of each moment; Calculate the The standard deviation of the time series data of nitrogen oxide gas concentration at the outlet of all SCR reactors within the target period at the moment ; At the injection flow rate Within the target period of time, , as the first The second-order difference of the ammonia injection flow rate at each moment, where: , as well as They are the ammonia injection flow rate Within the target period of time, moment, The moment and Ammonia injection flow value at a certain moment; In the Within the target period of time, The second-order difference of the ammonia injection flow rate at each moment is The ratio of The hysteresis significance of the ammonia injection flow rate at each moment; Get the The time corresponding to the local maximum value of the hysteresis significance of the ammonia injection flow rate at all times within the target period of a moment is used as the reference time; Get the For each target period at a reference time within a target period at each reference time, calculate the Pearson correlation coefficient between the nitrogen oxide gas concentration time series data at all SCR outlets and the ammonia injection flow rate within the target period at each reference time as the third correlation coefficient at each reference time, and obtain the largest of the third correlation coefficients at all reference times as the fourth maximum significant time lag value; The reference time corresponding to the maximum third correlation coefficient is recorded as The target moment corresponding to each moment; In the In the target period of each moment, the first The fourth time lag value corresponding to the moment.

6. The ammonia injection control method for the SCR denitration system based on the big data prediction model according to claim 5 is characterized in that: As stated in In the target period of each moment, the first The fourth time lag value corresponding to the moment includes the following specific steps: Get the target time period of all times The sum of the normalized values ​​of the lag significance at the reference moments; Get the fourth largest significant time lag value and the The ratio of the flue gas flow rate value at the target time corresponding to the first time is further obtained; the ratio is The product of the flue gas velocity values ​​at each moment; The first The sum of the normalized values ​​of the hysteresis significance of all reference moments in the target period of the moment, and the ratio of the The product of the flue gas velocity values ​​at the time is recorded as The fourth time lag value at this moment.

7. The SCR denitration system ammonia injection control method based on the big data prediction model according to claim 1 is characterized in that: The step of obtaining the second time lag value according to the fourth time lag value and the third time lag value comprises the following specific steps: The second time lag value is The fourth time lag value corresponding to the moment minus the The difference of the third time lag values ​​corresponding to the moments.

8. The ammonia injection control method for the SCR denitration system based on the big data prediction model according to claim 1 is characterized in that: The specific steps of regulating the ammonia injection flow rate according to the size of the second time lag value and the maximum ammonia flow rate when the valve is fully opened include the following: The second time lag is ; When the second time lag value is less than or equal to zero, use The nitrogen oxide gas concentration at the inlet of the SCR reactor at the time The nitrogen oxide gas concentration at the inlet of the SCR reactor at the moment is used to regulate the ammonia injection flow rate; When the second time lag value is greater than zero, the predicted gas value at the reactor inlet is obtained according to the power, and the valve opening percentage is obtained according to the predicted gas value at the reactor inlet and the maximum ammonia flow when the valve is fully opened.

9. The SCR denitration system ammonia injection control method based on the big data prediction model according to claim 8 is characterized in that: The specific steps of obtaining the predicted gas value at the reactor inlet according to the power are as follows: In the time series data of nitrogen oxide gas concentration at the inlet of the SCR reactor, As the The first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor at the moment, where and is the time series data of nitrogen oxide gas concentration at the inlet of SCR reactor The moment and The nitrogen oxide gas concentration value at the inlet of the SCR reactor at a certain moment; In power, As the The first-order difference of power at each moment is and The power The moment and The power value at a moment; Get At the moment, the first-order difference of the nitrogen oxide gas concentration time series data at the inlet of the SCR reactor is The ratio of the first-order difference of power at the time; further obtaining the mean of the ratio at all times, recorded as the influence coefficient; Build from 0 to The second time lag value sequence of is any value in the second time lag value sequence; The first time lag is ; Get the influence coefficient and At the moment, the product of the first-order difference of power is further obtained, and the product and At the time, the sum of the nitrogen oxide gas concentration values ​​at the inlet of the SCR reactor is recorded as the predicted sum value; Get the sum of all predicted values ​​for all values ​​in the second time lag value sequence, recorded as The predicted value of gas at the reactor inlet at the time .

10. The SCR denitration system ammonia injection control method based on the big data prediction model according to claim 8 is characterized in that: The specific steps of obtaining the valve opening percentage according to the predicted gas value at the reactor inlet and the maximum ammonia flow rate when the valve is fully opened include the following: According to the nitrogen oxide gas concentration data under the stable output state of the historical nitrogen oxide gas concentration, the corresponding optimal ammonia flow rate is obtained; Get pending parameters , The predicted value of gas at the reactor inlet at the time and the flue gas velocity Flue gas velocity value at the moment The product of the three; further according to the product of the three and the flow bias term The sum of the values, constructing and corresponding to the moment The equivalent relationship of the optimal ammonia flow rate under Based on the equivalent relationship and the optimal ammonia flow rate at all historical moments, the gas prediction value at the reactor inlet and the flue gas velocity Flue gas velocity value at the moment , perform function fitting and obtain and The value of is used to build a regression model; According to the regression model, the gas prediction value and flue gas flow rate value at the current moment are input to obtain the required ammonia flow at the current moment in real time; The percentage of the required ammonia flow rate at the current moment to the maximum ammonia flow rate when the valve is fully opened is recorded as the valve opening percentage at the current moment.

Images