Pollutant emission dynamic prediction and intelligent optimization method, system and device

Abstract

The present application relates to a pollution emission dynamic prediction and intelligent optimization method, system and device. The method comprises: periodically acquiring the latest window emission data through the sliding window mode, dynamically updating the real-time emission rate trend, and then performing the exponential moving average to obtain the smoothed emission rate trend; calculating the predicted total emission according to the cumulative actual emission and the smoothed emission rate trend; calculating the emission deviation value and the pollutant emission concentration required to be adjusted per hour; calculating the hourly allowed emission mean according to the adjustment factor, the emission deviation value and the pollutant emission concentration required to be adjusted per hour; calculating the feedforward control amount and the feedback control amount based on the predicted pollutant concentration and the hourly allowed emission mean; and calculating the final emission control amount through the dynamic weight allocation mode. The present application realizes precise control of flue gas pollutants and ensures that the total emission meets the standard within the control period.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental protection, in particular to a pollutant emission dynamic prediction and intelligent optimization method, system and device. BACKGROUND

[0002] The current control means of flue gas pollutant emission is divided into two categories, one is traditional PID control, which is based on the real-time monitoring of pollutant concentration data to construct a single feedback loop, and adjusts the operating parameters of desulfurization, denitrification and dust removal treatment equipment through proportional, integral and differential algorithm. However, PID control can only realize short-term closed-loop adjustment, and lacks the ability to predict and plan long-term emission targets. When the working condition fluctuates, the pollutant concentration is easy to exceed the standard due to the lag of parameter adjustment, and it is impossible to associate the cumulative value of the emission amount for forward-looking control; the second is artificial experience control, which relies on the on-site operator to manually adjust the equipment parameters according to historical experience and real-time data. This method is greatly affected by the operation level and experience difference of personnel, and has low control accuracy and slow response speed, and is difficult to cope with complex and variable working conditions, and cannot realize accurate calculation and dynamic regulation of long-term total emission. SUMMARY

[0003] Therefore, it is necessary to provide a pollutant emission dynamic prediction and intelligent optimization method, system and device which can realize accurate control of flue gas pollutants and ensure that the total emission amount meets the standard within the control period.

[0004] A pollutant emission dynamic prediction and intelligent optimization method, the method comprising:

[0005] determining a real-time emission rate trend, and constructing a linear regression model according to the real-time emission rate trend;

[0006] periodically obtaining the emission data of the latest window in a sliding window manner, and dynamically updating the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend;

[0007] performing exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend; calculating the predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend;

[0008] calculating the emission amount deviation value according to the total emission amount allowable value and the predicted total emission amount;

[0009] calculating the pollutant emission concentration that needs to be adjusted every hour according to the emission amount deviation value;

[0010] An adjustment factor is set, and the adjustment factor, the emission deviation value and the pollutant emission concentration required to be adjusted per hour are calculated to obtain an hourly allowed emission average value;

[0011] The LSTM model is trained through the emission data, and the LSTM model outputs a predicted pollutant concentration through training.

[0012] The feedforward control amount and the feedback control amount are calculated based on the predicted pollutant concentration and the hourly allowed emission average value.

[0013] Based on the feedforward control amount and the feedback control amount, the final emission control amount is calculated by dynamically allocating weights.

[0014] In another aspect, a pollutant emission dynamic prediction and intelligent optimization system is also provided, comprising:

[0015] A linear regression model construction module is configured to determine a real-time emission rate trend and construct a linear regression model based on the real-time emission rate trend.

[0016] An emission rate dynamic updating module is configured to periodically obtain emission data of a latest window through a sliding window method, and dynamically update the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend.

[0017] A total emission amount prediction module is configured to perform an exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend, and calculate a predicted total emission amount based on the cumulative actual emission amount and the smoothed emission rate trend.

[0018] An emission amount deviation value calculation module is configured to calculate an emission amount deviation value based on a total emission amount allowed value and the predicted total emission amount.

[0019] A required emission concentration adjustment calculation module is configured to calculate a pollutant emission concentration required to be adjusted per hour based on the emission amount deviation value.

[0020] An hourly allowed emission average value calculation module is configured to set an adjustment factor, and calculate an hourly allowed emission average value based on the adjustment factor, the emission amount deviation value and the pollutant emission concentration required to be adjusted per hour.

[0021] A model training module is configured to train an LSTM model through emission data, and output a predicted pollutant concentration through the trained LSTM model.

[0022] A feedforward control amount and a feedback control amount calculation module is configured to calculate a feedforward control amount and a feedback control amount based on the predicted pollutant concentration and the hourly allowed emission average value.

[0023] The emission control amount calculation module is configured to calculate a final emission control amount based on the feedforward control amount and the feedback control amount by dynamically allocating weights.

[0024] In another aspect, a computer device is also provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the pollution emission dynamic prediction and intelligent optimization method of any one of the above aspects when executing the computer program.

[0025] Compared with the prior art, the pollution emission dynamic prediction and intelligent optimization method, system and device provided by the present application have the following beneficial effects:

[0026] The present application periodically incorporates the latest emission data through the sliding window and the linear regression model, so that the calculation of the emission rate is always based on the real emission law, avoiding the prediction deviation caused by the long-term unchanged initial model parameters. The updated emission rate trend is subjected to the exponential moving average, which can effectively filter short-term random fluctuations. In combination with the cumulative actual emission amount, the predicted total emission amount is calculated, further reducing the gap between the predicted value and the true value, and improving the prediction accuracy. By calculating the emission deviation value, the pollutant emission concentration that needs to be adjusted per hour is derived, the target is quantified, and the dependence on artificial experience judgment is avoided, realizing precise emission control. By introducing the adjustment factor to dynamically adjust the hourly allowable emission mean value, the flexibility and adaptability are improved. The LSTM model is used to obtain the predicted pollutant concentration, and the feedforward control amount and the feedback control amount are combined to realize precise control of the hourly mean value.

[0027] The present application can realize precise control of flue gas pollutants, optimize the operation cost while ensuring that the total emission amount meets the standard within the control period. It is particularly suitable for complex emission scenes, large fluctuations in working conditions, and long-term target control in the field of industrial enterprises, providing a practical technical solution for precise and scientific pollution control of pollutants, and improving the fine management level of emission control. BRIEF DESCRIPTION OF DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and those skilled in the art can obtain other related drawings according to these drawings without creative labor.

[0029] Figure 1 The flowchart of the pollution emission dynamic prediction and intelligent optimization method provided in embodiment 1 is shown in the figure.

[0030] Figure 2 The structural block diagram of the pollution emission dynamic prediction and intelligent optimization system provided in embodiment 2 is shown in the figure.

[0031] Figure 3 Figure 1 is a schematic diagram of the internal structure of the computer device provided in Embodiment 3.

[0032] The purposes, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work are within the protection scope of the present application.

[0034] It can be understood that the technical solutions in each embodiment of the present application can be combined with each other, but it must be based on the fact that a person of ordinary skill in the art can realize the combination, and when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope of the present application.

[0035] The embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application.

[0036] Embodiment 1

[0037] As shown in FIG. 1, the present embodiment provides a dynamic prediction and intelligent optimization method for pollutant emission, which comprises the following steps: Figure 1 Step 201, determining a real-time emission rate trend, and constructing a linear regression model according to the real-time emission rate trend.

[0038] Step 202, periodically acquiring the emission data of the latest window in a sliding window manner, and dynamically updating the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend.

[0039] Step 203, performing an exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend, and calculating a predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend.

[0040] Step 204, calculating an emission amount deviation value according to the total emission amount allowable value and the predicted total emission amount.

[0041] Step 205, calculating a pollutant emission concentration that needs to be adjusted every hour according to the emission amount deviation value.

[0042] Step 205, calculating a pollutant emission concentration that needs to be adjusted every hour according to the emission amount deviation value.

[0043] Step 206, set the adjustment factor, calculate the hourly allowable emission mean according to the adjustment factor, the emission deviation value and the pollutant emission concentration required to be adjusted per hour.

[0044] Step 207, train the LSTM model through the emission data, and then output the predicted pollutant concentration through the trained LSTM model.

[0045] Step 208, calculate the feedforward control amount and the feedback control amount based on the predicted pollutant concentration and the hourly allowable emission mean.

[0046] Step 209, based on the feedforward control amount and the feedback control amount, calculate the final emission control amount by dynamically allocating weights.

[0047] In the specific implementation process of step 201, first, the emission data is collected, and the collected emission data needs to cover five dimensions including flue gas characteristics, reaction conditions, denitration parameters, desulfurization parameters and emission indicators, and the collection of specific parameters for different processes is strengthened. As shown in Table 1, the parameter diagram of the emission data is given.

[0048] Table 1: Emission data parameters

[0049]

[0050] Among them, the sensor collects the running data of the equipment in real time; the edge gateway runs the edge computing algorithm to filter and compress the data; the OPC UA server receives the edge gateway data and provides a standardized interface. The collected data is stored in the database for later calling. Other input parameters, such as control targets, control periods and production plans, are manually entered. It is worth mentioning that in addition to online collection of real-time data through sensors, the data collection method also includes offline sampling and laboratory analysis data.

[0051] Then, determine the window time length Select the emission data of the latest window time length to build a linear regression model.

[0052] In this embodiment, the latest hour (60 minutes) of emission data is selected, and one point is sampled per minute, based on which a linear regression model is built, which is expressed as:

[0053] ;

[0054] In the formula, represents the emission rate corresponding to the time period; represents the emission rate trend; represents the One time period; This represents the intercept term.

[0055] In the specific implementation of step 202, the interception time period is set to... And determine the window time length. Then, by using a sliding window approach, the emission data of the latest window is periodically acquired. In this embodiment, the time period is set to 5 minutes. By acquiring the emission data of the latest window every 5 minutes (e.g., if the window time length is set to 60 minutes, one data point is acquired per second), the real-time emission rate trend is dynamically updated based on a linear regression model to obtain the updated emission rate trend, expressed as:

[0056] ;

[0057] In the formula, This indicates the updated emission rate trend; The amount of data representing emission data is taken in this embodiment as... .

[0058] In the specific implementation of step 203, an exponential moving average is applied to the updated emission rate trend to obtain a smoothed emission rate trend, expressed as:

[0059] ;

[0060] In the formula, This indicates the smoothed emission rate trend over the current time period; Indicates the smoothing factor; This represents the smoothed emission rate trend after the previous time period. It's worth noting that the smoothing factor... Control the weighting of historical data and current data. The larger the value, the more sensitive the model is to the latest data (faster response but greater fluctuations); The smaller the value, the smoother the model (slower response but more stable).

[0061] Then, the cumulative actual emissions are obtained every 5 minutes. Based on the cumulative actual emissions and the smoothed emission rate trend, the predicted total emissions are calculated, as expressed by:

[0062] ;

[0063] In the formula, This represents the projected total emissions within the target period; This indicates the cumulative actual emissions; This indicates the remaining time of the target period.

[0064] In the implementation of step 204, the emission deviation value is calculated according to the total emission allowance value and the predicted total emission value, and the expression is as follows:

[0065] ;

[0066] In the expression, the emission deviation value is represented by ; The total emission allowance value is represented by The total emission allowance value is the total amount of pollutants allowed to be emitted in the target period, which is a known value.

[0067] In the implementation of step 205, the pollutant emission concentration that needs to be adjusted per hour is calculated according to the emission deviation value, and the expression is as follows:

[0068] ;

[0069] In the expression, the pollutant emission concentration that needs to be adjusted per hour is represented by ; The emission deviation value is represented by ; The remaining time of the target period is represented by The average flue gas flow value in the target period is represented by

[0070] In the implementation of step 206, in order to avoid excessive adjustment and achieve smooth fluctuations, an adjustment factor is introduced, and the hourly allowed emission average is calculated according to the adjustment factor, the emission deviation value, and the pollutant emission concentration that needs to be adjusted per hour. The expression of the adjustment factor is as follows:

[0071] K= ;

[0072] The calculation expression of the hourly allowed emission average is as follows:

[0073] ;

[0074] In the expression, K represents the adjustment factor; The over-emission penalty coefficient is represented by , which is a coefficient greater than 1. The larger the value, the smaller the system allows to relax; The energy-saving reward coefficient is represented by , which is a coefficient between 0 and 1. The larger the value, the greater the intensity of reduction; The total emission allowance value of the pollutants in the target period is represented by ; The hourly allowed emission average is represented by ; The current hourly average actual value is represented by .

[0075] In the implementation of step 207, in order to predict the pollutant concentration change in the future period based on historical data, a compensation signal is generated in advance to compensate for the time lag effect of the desulfurization and denitrification process. Through the fusion of LSTM feedforward and PID feedback correction, an optimized control instruction is output.

[0076] Specifically, emission data is obtained, and the LSTM model is trained by the emission data. However, in order to ensure the accuracy of the model, input data with weak correlation with the output value needs to be screened out and removed, and the influence factor with relatively large correlation needs to be determined, and the LSTM model is trained by the influence factor. The screening method adopts the mutual information rule. Mutual information is an entropy in information theory for measuring information. The size represents the mutual relationship between two variables. The larger the value is, the more complex the relationship between the two variables is. Conversely, the smaller the value is, the larger the value of the influence factor is reserved as the data set for model training.

[0077] Further, in order to ensure high-quality input data of the model, the data needs to be preprocessed. The steps are as follows: align the original data set to eliminate invalid data, unify the time stamp, and ensure data integrity; then generate lag features, build features reflecting process lag effects, so that the model can capture causal time series relationships; then perform sliding window statistics to extract time series dynamic patterns such as trends and volatility; then perform standardization / encoding to eliminate dimension differences, adapt to model input requirements, and complete data preprocessing. Finally, the data is divided into training set, validation set and test set, and the division ratio is set conventionally and will not be described here. It is worth noting that in addition to the above methods, missing value processing, outlier detection and processing, normalization and standardization, logarithmic transformation, feature construction, smoothing processing, etc. can also be used for data preprocessing, which will not be described here.

[0078] When training the model, the input data set includes time series data (sliding window 30 minutes, sampling interval 1 minute), pollutant concentration (NOx, SO2), control quantity (ammonia injection amount, desulfurizing agent flow), working condition parameter (flue gas flow, temperature, oxygen content), etc. Based on the set mean square error (MSE) loss function and prediction bias penalty term, the model is trained through online learning, and finally a trained LSTM model is obtained, which outputs the predicted pollutant concentration. The online learning method is to incrementally update the model parameters every 24 hours to adapt to the working condition drift. The prediction of the pollutant concentration time is not long and can be adjusted according to the actual situation. In this embodiment, the output is the predicted value of the pollutant concentration in the next 5-10 minutes.

[0079] In the implementation of step 208, the feedforward and feedback collaborative control is performed, in which the feedforward gain compensates for the time lag and disturbance in advance using the prediction information, and the PID feedback handles the remaining error.

[0080] wherein the feedforward control amount calculation expression is:

[0081] ;

[0082] ;

[0083] wherein, represents the feedforward control amount; represents the feedforward gain; represents the predicted pollutant concentration; represents the data amount of the control amount; represents the control amount corresponding to the th time period; represents the control amount mean value; represents the pollutant concentration variation corresponding to the th time period; represents the pollutant concentration variation mean value.

[0084] The feedback control amount calculation expression is:

[0085] ;

[0086] ;

[0087] wherein, represents the feedback control amount; represents the proportional coefficient, used to control the response strength of the control action to the current error; represents the integral time, used to eliminate the speed of the steady-state error (the smaller Ti is, the stronger the integral action is); represents the derivative time, used to predict the advance of the future trend; represents the actual pollutant concentration at the current time; represents the deviation value at the current time; represents the hourly allowable emission mean value at the current time; represents the deviation value at the integral time variable τ; represents the lag time, i.e. the time from the start of the control action to the change of the feedback data.

[0088] wherein the proportional coefficient , the integral time , and the derivative time are calculated according to the following expressions:

[0089] ;

[0090] ;

[0091] ;

[0092] In the formula, represents the system gain, i.e. the static amplification multiple of the change of the input quantity of the system to the change of the output quantity; represents the time constant, i.e. the time for the system to reach a steady state.

[0093] In the implementation process of step 209, according to the deviation degree of the current concentration from the center of the target interval, a weight is dynamically allocated through a formula, so as to avoid the continuous deviation of the emission concentration from the target interval, and ensure that the pollutant emission concentration always converges to the target interval stably.

[0094] The dynamically allocated weight is The expression of the dynamically allocated weight is:

[0095]

[0096] The expression of the final emission control quantity is calculated as:

[0097]

[0098] In the formula, represents the final emission control quantity; represents the maximum feedforward weight; represents a key real-time state variable, which is determined according to the field working condition, such as the inlet pollutant concentration or the boiler load or the coal supply amount, etc.; represents the midpoint of the target interval; represents the half-width of the interval.

[0099] Finally, the optimization instruction (such as the ammonia injection amount, the desulfurizing agent injection amount) is generated based on the final emission control quantity, and the optimization instruction is converted into a physical action through an execution output module. It can be understood that the execution output module is the end execution unit of the desulfurization and denitrification control system, which is composed of a signal analysis, a driving circuit, an execution mechanism (a high-precision regulating valve, a variable frequency feeder) and a feedback unit. It receives the optimization instruction in real time, drives the valve opening or the motor speed, and dynamically adjusts the environmental protection reagent injection amount.

[0100] The present application can automatically trigger the intensive control mode during the orange / red early warning weather, dynamically tighten the emission limit value, reduce the pollution load in advance, and realize the emission standard. Through intelligent optimization control, the over-addition of environmental protection reagents (ammonia water, limestone, etc.) is reduced, and the energy consumption and material waste are reduced. The method of the present application can effectively reduce the labor intensity and reduce the decision-making errors.

[0101] It should be understood that, although the present embodiment Figure 1 ​​The steps in each of the above processes are displayed in sequence according to the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least a part of the steps in each of the above processes can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be alternately executed with at least a part of other steps or sub-steps or stages of other steps.

[0102] Embodiment 2

[0103] Based on the pollutant emission dynamic prediction and intelligent optimization method in Embodiment 1, this embodiment discloses a pollutant emission dynamic prediction and intelligent optimization system, as shown in Figure 2 The pollutant emission dynamic prediction and intelligent optimization system includes a linear regression model construction module 401, an emission rate dynamic updating module 402, a total emission amount prediction module 403, an emission amount deviation value calculation module 404, a required adjustment emission concentration calculation module 405, an hourly allowed emission average value calculation module 406, a model training module 407, a feedforward control amount and feedback control amount calculation module 408, and an emission control amount calculation module 409, wherein:

[0104] The linear regression model construction module 401 is used to determine the real-time emission rate trend and construct a linear regression model according to the real-time emission rate trend.

[0105] The emission rate dynamic updating module 402 is used to periodically obtain the latest window of emission data in a sliding window manner, and dynamically update the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend.

[0106] The total emission amount prediction module 403 is used to perform an exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend, and calculate the predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend.

[0107] The emission amount deviation value calculation module 404 is used to calculate the emission amount deviation value according to the total emission amount allowed value and the predicted total emission amount.

[0108] The required adjustment emission concentration calculation module 405 is used to calculate the pollutant emission concentration that needs to be adjusted per hour according to the emission amount deviation value.

[0109] The hourly allowable average emission calculation module 406 is used to set the adjustment factor. Based on the adjustment factor, the emission deviation value and the pollutant emission concentration that needs to be adjusted per hour, the hourly allowable average emission is calculated.

[0110] The model training module 407 is used to train an LSTM model using emission data, and then output the predicted pollutant concentrations using the trained LSTM model.

[0111] The feedforward control quantity and feedback control quantity calculation module 408 is used to calculate the feedforward control quantity and feedback control quantity based on the predicted pollutant concentration and the hourly allowable average emission.

[0112] The emission control quantity calculation module 409 is used to calculate the final emission control quantity based on the feedforward control quantity and the feedback control quantity by dynamically allocating weights.

[0113] In this embodiment, the specific working processes and principles of the linear regression model construction module 401, emission rate dynamic update module 402, total emission prediction module 403, emission deviation calculation module 404, emission concentration adjustment calculation module 405, hourly allowable average emission calculation module 406, model training module 407, feedforward control quantity and feedback control quantity calculation module 408, and emission control quantity calculation module 409 are the same as those in Embodiment 1, and therefore will not be described again in this embodiment. Each unit module can be implemented entirely or partially through software, hardware, or a combination thereof. Each unit module can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each of the above unit modules.

[0114] Example 3

[0115] like Figure 3 The diagram illustrates a terminal device disclosed in this embodiment, comprising a transmitter, a receiver, a memory, and a processor. The transmitter transmits instructions and data, the receiver receives instructions and data, the memory stores computer-executed instructions, and the processor executes the computer-executed instructions stored in the memory to implement the method described in Embodiment 1 above.

[0116] It is important to note that the aforementioned memory can be either standalone or integrated with the processor. When the memory is set up independently, the terminal device also includes a bus for connecting the memory and the processor.

[0117] Example 4

[0118] The embodiment discloses a computer readable storage medium, and computer readable storage medium has computer execution instructions stored therein. When a processor executes the computer execution instructions, the method in the above embodiment 1 is realized.

[0119] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the computer program can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM).

[0120] The technical features of the above embodiments can be combined in any manner. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combinations of the technical features do not contradict, they should be considered within the scope of the present disclosure.

[0121] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be pointed out that, for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the protection scope of the present application.

Claims

1. A dynamic prediction and intelligent optimization method for pollutant emission, characterized in that, The method comprises: determining a real-time emission rate trend, and constructing a linear regression model according to the real-time emission rate trend; periodically acquiring emission data of a latest window in a sliding window manner, and dynamically updating the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend; performing exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend; and calculating a predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend; calculating an emission amount deviation value according to the total emission amount and the predicted total emission amount; calculating a pollutant emission concentration that needs to be adjusted per hour according to the emission amount deviation value; setting an adjustment factor, and calculating a per-hour allowed emission average value according to the adjustment factor, the emission amount deviation value, and the pollutant emission concentration that needs to be adjusted per hour; training an LSTM model based on the emission data, and then outputting a predicted pollutant concentration by using the trained LSTM model; calculating a feedforward control amount and a feedback control amount based on the predicted pollutant concentration and the per-hour allowed emission average value; calculating a final emission control amount by dynamically allocating weights based on the feedforward control amount and the feedback control amount; setting an adjustment factor, and calculating a per-hour allowed emission average value according to the adjustment factor, the emission amount deviation value, and the pollutant emission concentration that needs to be adjusted per hour, comprising: setting the adjustment factor, which is expressed as: K= ； calculating the per-hour allowed emission average value according to the adjustment factor, the emission amount deviation value, and the pollutant emission concentration that needs to be adjusted per hour, and the calculation expression is: ； In the formula, K represents an adjustment factor; represents an over-discharge penalty coefficient; represents an energy-saving reward coefficient; represents a total amount of pollutant discharge allowed in a target period; represents an hourly average allowed emission value; represents a current hourly average actual value; represents a pollutant discharge concentration that needs to be adjusted per hour; represents an emission deviation value.

2. The method of dynamic prediction and intelligent optimization of pollutant emission amount according to claim 1, characterized in that, dynamically updating the real-time emission rate trend based on the linear regression model to obtain the updated emission rate trend, comprising: the constructed linear regression model is expressed as: ； dynamically updating the real-time emission rate trend based on the linear regression model to obtain the updated emission rate trend, and the expression is: ； wherein represents the emission rate corresponding to the th time period; represents the emission rate trend; represents the intercept term; represents the updated emission rate trend; represents the amount of data involved in the calculation; represents the emission rate corresponding to the th time period.

3. The method of dynamic prediction and intelligent optimization of pollutant emission quantity according to claim 2, characterized in that, performing exponential moving average on the updated emission rate trend to obtain the smoothed emission rate trend, and the expression is: ； wherein represents the smoothed emission rate trend for the current time period; represents the smoothing factor; represents the smoothed emission rate trend for the previous time period.

4. The method of dynamic prediction and intelligent optimization of pollutant emission amount according to claim 3, characterized in that, calculating the predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend, and the expression is: ； In the formula, represents the predicted total emission amount within the target period; represents the cumulative actual emission amount; represents the remaining time of the target period.

5. The method of dynamic prediction and intelligent optimization of pollutant emission amount according to any one of claims 1 to 4, characterized in that, calculating the pollutant emission concentration that needs to be adjusted per hour according to the emission amount deviation value, and the expression is: ； In the formula, represents the target cycle remaining time; represents the average flue gas flow value in the target cycle.

6. The method of dynamic prediction and intelligent optimization of pollutant emission quantity according to claim 5, characterized in that, calculating the feedforward control amount and the feedback control amount based on the predicted pollutant concentration and the per-hour allowed emission average value, comprising: calculating the feedforward control amount, and the expression is: ； calculating the feedback control amount, and the expression is: ； ； In the formula, represents a feedforward control amount; represents a feedforward gain; represents a predicted pollutant concentration; represents a feedback control amount; represents a proportional coefficient; represents an integral time; represents a derivative time; represents an actual pollutant concentration at the current time; represents a deviation value at the current time; represents an hourly allowable emission mean value at the current time; represents a deviation value at the integral time variable τ; represents a lag time.

7. The method of dynamic prediction and intelligent optimization of pollutant emission amount according to claim 6, characterized in that, calculating the final emission control amount by dynamically allocating weights based on the feedforward control amount and the feedback control amount, and the expression is: ； wherein the dynamic allocation weight is expressed by the expression: ； wherein represents the final emission control amount; represents the maximum feedforward weight; represents the key real-time state variable; represents the target interval midpoint; represents the interval half-width.

8. A dynamic prediction and intelligent optimization system for pollutant emission, characterized in that, The system comprises: a linear regression model construction module, configured to determine a real-time emission rate trend, and construct a linear regression model according to the real-time emission rate trend; an emission rate dynamic updating module, configured to periodically acquire emission data of a latest window in a sliding window manner, and dynamically update the real-time emission rate trend based on the linear regression model to obtain an updated emission rate trend; The total emission amount prediction module is configured to perform exponential moving average on the updated emission rate trend to obtain a smoothed emission rate trend; and calculate the predicted total emission amount according to the cumulative actual emission amount and the smoothed emission rate trend. The emission amount deviation value calculation module is configured to calculate the emission amount deviation value according to the total emission amount allowable value and the predicted total emission amount. The emission concentration adjustment calculation module is configured to calculate the pollutant emission concentration that needs to be adjusted per hour according to the emission amount deviation value. The hourly allowable emission average calculation module is configured to set an adjustment factor, and calculate the hourly allowable emission average according to the adjustment factor, the emission amount deviation value and the pollutant emission concentration that needs to be adjusted per hour. The model training module is configured to train an LSTM model through emission data, and then output a predicted pollutant concentration through the trained LSTM model. The feedforward control amount and feedback control amount calculation module is configured to calculate a feedforward control amount and a feedback control amount based on the predicted pollutant concentration and the hourly allowable emission average. The emission control amount calculation module is configured to calculate a final emission control amount based on the feedforward control amount and the feedback control amount through a dynamic weight allocation manner. In the hourly allowable emission average calculation module, the adjustment factor is set, and the hourly allowable emission average is calculated according to the adjustment factor, the emission amount deviation value and the pollutant emission concentration that needs to be adjusted per hour. The adjustment factor is set and expressed as follows: K= ； The hourly allowable emission average is calculated according to the adjustment factor, the emission amount deviation value and the pollutant emission concentration that needs to be adjusted per hour, and the calculation expression is as follows: ； In the formula, K represents an adjustment factor; represents an over-discharge penalty coefficient; represents an energy-saving reward coefficient; represents a total amount of pollutant emissions allowed in a target period; represents an hourly average allowed emission value; represents a current hourly average actual value; represents a pollutant emission concentration that needs to be adjusted per hour; represents an emission deviation value. 9.A computer device, comprising a memory and a processor, and characterized in that, The memory stores a computer program, and the processor implements the steps of the pollutant emission amount dynamic prediction and intelligent optimization method according to any one of claims 1 to 7 when executing the computer program.

Images