Pulverized coal horizontal boiler detection control system and use method

Abstract

The invention belongs to the technical field of combustion equipment control, and particularly discloses a pulverized coal horizontal boiler detection control system and a use method. The system disclosed by the invention comprises a data acquisition module, a data processing module, a prediction model module, an optimization algorithm module and a system execution and display module. The data acquisition module is used for acquiring operation data of the boiler in real time; the data processing module is used for preprocessing the operation data; the prediction model module is used for outputting predicted values of nitrogen oxide and ammonia gas emission concentration in a future period; the optimization algorithm module is used for calculating optimized boiler operation parameters on line by adopting a particle swarm optimization algorithm; and the system execution and display module is used for controlling the boiler to operate according to the optimized operating parameters and displaying the operating state of the boiler in real time. The intelligent control level and the combustion efficiency of coal and ammonia co-combustion can be improved, and pollutant emission is reduced. The method can be widely applied to coal and ammonia co-combustion control of the pulverized coal boiler.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of combustion equipment control, and particularly relates to a coal powder horizontal boiler detection control system and a use method. BACKGROUND

[0002] Under the background of the national strategic goal of carbon peak and carbon neutrality, optimization of energy structure and industrial emission reduction are in urgent need. As the main equipment for power production and industrial energy supply, coal-fired boilers have particularly prominent problems of carbon emissions and pollutant emissions. According to statistics, global industrial boilers consume about 4 billion tons of standard coal per year, accounting for more than 35% of global energy consumption, of which CO2 emissions from coal combustion account for more than 40% of energy-related carbon emissions, and NO x , SO2 and other pollutants also cause serious environmental problems.

[0003] Ammonia (NH3) as a zero-carbon fuel has the advantages of convenient storage and transportation, and mature industry, and is an important choice for low-carbon transformation of coal combustion. However, the emission control of NO x and NH3 in the process of coal-ammonia mixed combustion is extremely complex, and the traditional control method relies on experience adjustment, which is difficult to realize real-time optimization, and has problems such as low control precision, response lag and insufficient intelligence.

[0004] In the prior art, there are some emission prediction methods based on machine learning models, such as a NO x prediction method based on LSSVM (CN109508818A), but it mostly uses offline training and sliding window updating, which is difficult to adapt to the dynamic changes of the combustion process; there is also a pump station monitoring system based on Internet of Things (CN120211375A), which uses MQTT protocol to realize remote data transmission, but it does not involve modeling and optimization of the combustion process.

[0005] As a neural network suitable for time series data, long short-term memory network (LSTM) has advantages in dynamic prediction, but its combination with optimization algorithms is still less studied in boiler combustion control, and there are problems such as difficulty in model and algorithm cooperation and insufficient real-time performance.

[0006] Therefore, there is an urgent need for a system that can realize real-time monitoring, intelligent prediction and multi-objective optimization of the coal-ammonia mixed combustion process to improve combustion efficiency, reduce pollutant emissions and promote the intelligent upgrading of industrial combustion equipment. SUMMARY

[0007] The purpose of the present application is to provide a coal powder horizontal boiler detection control system that can realize real-time monitoring of key parameters in the combustion process, intelligent prediction of NO xAccurate prediction of NH3 emission concentration, online collaborative optimization of operation parameters, and integrated management and visual display of data, thereby improving the intelligent control level of coal-ammonia mixed combustion, combustion efficiency and reducing pollutant emissions. The second object of the present application is to provide a method for using the above system.

[0008] To achieve the above objects, the technical solutions adopted by the present application are as follows: A coal powder horizontal boiler detection and control system, comprising a data acquisition module, a data processing module, a prediction model module, an optimization algorithm module, and a system execution and display module. The data acquisition module is used to acquire real-time operation data of the boiler and transmit the operation data to the data processing module. The data processing module is used to preprocess the operation data to obtain standardized time series data and transmit the standardized time series data to the prediction model module. The prediction model module has a neural network model trained based on a long short-term memory network, is used to receive the standardized time series data, output predicted values of nitrogen oxide and ammonia emission concentrations in a future period, and transmit the predicted values to the optimization algorithm module. The optimization algorithm module is used to take the predicted values as one of the optimization targets, calculate the optimized boiler operation parameters online using a particle swarm optimization algorithm, and transmit the optimized operation parameters to the system execution module. The system execution and display module controls the boiler to operate according to the optimized operation parameters and displays the boiler operation state in real time.

[0009] As a limitation, the operation data of the boiler includes gas flow, primary air volume, secondary air volume, flue gas oxygen content, furnace temperature, instantaneous coal supply amount, instantaneous ammonia flow, and NOx concentration in flue gas. x NH3 concentration.

[0010] As a second limitation, the preprocessing in the data processing module includes removal of outliers, data format conversion, standardization, and serialization according to time windows.

[0011] As a third limitation, the feature extraction layer in the neural network structure of the prediction model module is two Conv1D layers and two MaxPooling1D layers. The time series processing layer is two LSTM layers, and the output layer is two Dense layers.

[0012] As a fourth limitation, the fitness function constructed by the particle swarm optimization algorithm is the sum of the negative weighting of nitrogen oxide emission concentration and the positive weighting of combustion efficiency.

[0013] As a fifth limitation, the optimized operating parameters calculated by the particle swarm optimization algorithm include the ammonia-coal ratio, the primary air flow, and the secondary air door opening.

[0014] As a sixth limitation, the system executes and displays data in real time, including ammonia instantaneous flow, coal instantaneous flow, air instantaneous flow, combustion air oxygen content, and furnace temperature.

[0015] As a seventh limitation, the system execution and display module further includes a database unit for unified storage and management of historical data, real-time data, model parameters, and user information.

[0016] A method for using a coal powder horizontal boiler detection and control system, which uses the above-mentioned coal powder horizontal boiler detection and control system, includes the following steps in sequence: S1, real-time acquisition of boiler operating data; S2, preprocessing the operating data to obtain standardized time series data; S3, inputting the standardized time series data into a pre-trained LSTM prediction model to obtain the predicted values of NO x and NH3 emission concentration; S4, the optimization algorithm module uses the particle swarm optimization algorithm to solve the optimized boiler operating parameters online with the predicted emission concentration and calculated combustion efficiency as the target; S5, the optimized operating parameters are sent to the field gateway device via MQTT communication and written into the PLC controller to adjust the boiler operation; S6, the display screen is updated in real time to show the system operation state, data and optimization results, and new data is stored in the database for model updating.

[0017] The present application has the following technical advantages compared with the prior art: (1) The system of the present application can accurately predict the emission concentrations of nitrogen oxides and ammonia by real-time data acquisition and standardization, combined with an LSTM neural network model, overcoming the poor dynamic adaptability of traditional methods. The particle swarm optimization algorithm is used to adjust the operating parameters online, achieving multi-objective optimization of combustion efficiency and emission control, and solving the response lag problem caused by experience adjustment. Finally, the system execution module accurately regulates the boiler operation, significantly improves the combustion efficiency and reduces pollutant emissions, and enhances the operation transparency with the real-time display function, promoting the intelligentization and low-carbonization of industrial combustion equipment, providing an efficient solution for energy structure optimization and industrial emission reduction. (2) The system of the present invention adopts a neural network model trained by a long short-term memory network, and realizes multi-scale combustion feature extraction through dual Conv1D layers. It combines dual MaxPooling1D layers for feature dimensionality reduction and key information screening, which effectively improves the adaptability to dynamic combustion processes. Dual LSTM layers further capture long-term time-series dependencies, solving the problem that traditional methods cannot handle nonlinear coupling of combustion parameters. Finally, dual Dense layers output NO respectively. x The model, along with NH3 concentration prediction values, achieves high-precision multi-objective emission prediction, significantly enhancing the model's response to the complex dynamics of coal-ammonia co-combustion. This provides more reliable input for optimization algorithms, thereby supporting real-time collaborative optimization of combustion efficiency and emission control, and promoting the transformation of boiler systems towards intelligent and low-carbon operation. (3) The system of the present invention constructs a fitness function by combining negative weighting of NOx emission concentration with positive weighting of combustion efficiency, so that the particle swarm optimization algorithm automatically balances emission reduction and energy efficiency targets during the iteration process, avoiding parameter imbalance caused by multi-objective conflicts in traditional methods; at the same time, it optimizes key operating parameters such as ammonia-to-coal ratio, primary air volume and secondary damper opening, which can accurately control the combustion reaction rate and air mixing degree, and effectively suppress NO x It generates and improves the combustion efficiency of ammonia, enhances the real-time response capability of the coal-ammonia co-combustion system, and ensures stable boiler operation while reducing pollutant emissions, providing an efficient solution for the intelligent and low-carbon transformation of industrial combustion equipment. (4) The system of the present invention provides operators with an intuitive combustion status monitoring interface by displaying key parameters such as ammonia, pulverized coal, air flow and combustion air oxygen content, and furnace temperature in real time, which facilitates quick identification of abnormalities and adjustment of operating strategies. At the same time, the added database unit realizes unified storage and management of historical data, real-time data, model parameters and user information, which not only supports the continuous optimization and parameter iteration of emission prediction models, but also provides data support for fault diagnosis, energy efficiency analysis and equipment maintenance.

[0018] This invention belongs to the field of combustion equipment control technology and can realize real-time monitoring of key parameters of the combustion process, NO. x With accurate prediction of NH3 emission concentration, online collaborative optimization of operating parameters, and integrated management and visualization of data, the intelligent control level of coal-ammonia co-combustion, combustion efficiency, and pollutant emissions can be improved. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0020] In the attached diagram: Figure 1This is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a diagram of the LSTM neuron structure in Embodiment 1 of the present invention; Figure 3 This is a diagram of the dynamic neural network structure of Embodiment 1 of the present invention; Figure 4 This is a flowchart of the processing in Embodiment 2 of the present invention. Detailed Implementation

[0021] The preferred embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention.

[0022] Example 1 like Figure 1 As shown, this embodiment is a detection and control system for a pulverized coal horizontal boiler, including a data acquisition module, a data processing module, a prediction model module, an optimization algorithm module, and a system execution and display module.

[0023] The boiler used in this embodiment is a pulverized coal horizontal boiler, including the combustion furnace body, pulverized coal main burner, pulverized coal preparation and feeding system, natural gas / ammonia pipeline, combustion air fan, air preheater, exhaust gas purification system, induced draft fan, valve group, instrumentation, and control cabinet. The boiler body is a cylindrical horizontal structure with a diameter of 2m and a length of 8m. The effective volume of the combustion chamber is 12.3m³, the rated thermal power is 1MW, the rated pulverized coal combustion capacity of a single unit is 176kg / h, and the maximum consumption is 300kg / h. The pulverized coal conveying system is set up outdoors in a modified 20GP container house, which houses the silo, feeder, primary air fan, pulverized coal pipeline, ventilation, and fire extinguishing system.

[0024] The data acquisition module is used to collect boiler operating data in real time and transmit the data to the data processing module. In this embodiment, operating data is collected by sensors deployed on the boiler and a Siemens S7-1200 PLC, and transmitted to the cloud platform via a USR-PLCNET210 gateway and MQTT protocol. The transmitted data includes gas flow rate, primary air volume, secondary air volume, flue gas oxygen content, furnace temperature, instantaneous coal feed rate, instantaneous ammonia flow rate, and NO content in the flue gas. x With NH3 concentration.

[0025] The data processing module is used to preprocess the running data to obtain standardized time series data, and then pass the standardized time series data to the prediction model module.

[0026] Specifically, preprocessing includes outlier removal, data format conversion, standardization, and serialization by time window, where the standardization uses the Z-score method.

[0027] The prediction model module, with a built-in neural network model trained on a Long Short-Term Memory (LSTM) network, receives the standardized time-series data, outputs predicted values ​​for nitrogen oxide and ammonia emission concentrations for future periods, and transmits these predicted values ​​to the optimization algorithm module. Its LSTM neuron structure is as follows: Figure 2 As shown.

[0028] Specifically, the neural network model was compiled in Python 3.9 and built using the Tensorflow platform. The selected input features for the neural network were ammonia flow rate, primary air volume, furnace temperature, flue gas oxygen content, and coal feed rate. The stronger the correlation between the feature values ​​and the target, the higher the accuracy of the corresponding model training. The model was constructed using Sequential, retaining the dynamic feature processing structure, namely two Conv1D layers and two MaxPooling1D layers for feature extraction, two LSTM layers for processing time-series data, and two Dense layers as output layers. Its neural network structure diagram is shown below. Figure 3 As shown. The model is trained using historical data, with mean squared error as the loss function.

[0029] The optimization algorithm module is used to calculate the optimized boiler operating parameters online using the predicted value as one of the optimization objectives, and then transmit the optimized operating parameters to the system execution module.

[0030] In this embodiment, a particle swarm optimization algorithm is used. The input parameters are the ammonia-to-coal ratio (0-30%), primary air volume (800-1200 m³ / h), and secondary damper opening. The constructed fitness function is NO. x The weighted combination of emission concentration and combustion efficiency, where NO x The weight for emissions is -0.7 (minimize emissions), and the weight for combustion efficiency is 0.3 (maximize efficiency).

[0031] The system execution and display module controls the boiler to operate according to the optimized operating parameters and displays the boiler's operating status in real time.

[0032] Specifically, a MySQL database is used for data management, providing a unified storage and management system for historical data, real-time data, model parameters, and user information. Optimized operating parameters are published to a designated topic via MQTT. The gateway subscribes to this topic and receives instructions, which are then written to the corresponding registers of the PLC via the S7 protocol, enabling automatic parameter adjustment.

[0033] In addition, the system execution and display module is equipped with a display screen that can display the boiler's operating parameters and graphs in real time, including instantaneous ammonia flow rate, instantaneous pulverized coal flow rate, instantaneous air flow rate, combustion air oxygen content, furnace temperature, ammonia-to-coal ratio (circular gauge), and NO. xComparison of predicted and measured curves and alarm logs for exceeding limits.

[0034] In summary, this embodiment enables real-time monitoring of key parameters in the combustion process, including NO. x With accurate prediction of NH3 emission concentration, online collaborative optimization of operating parameters, and integrated management and visualization of data, the intelligent control level of coal-ammonia co-combustion, combustion efficiency, and pollutant emissions can be improved.

[0035] Example 2 This embodiment describes the usage method of a detection and control system for a pulverized coal horizontal boiler, as described in Embodiment 1. Figure 4 As shown, the steps are performed sequentially: S1. Real-time acquisition of boiler operating data.

[0036] S2. Preprocess the running data to obtain standardized time series data.

[0037] S3. Input the standardized time series data into the pre-trained LSTM prediction model to obtain the future time period NO. x Compared with the predicted NH3 emission concentration.

[0038] S4. The optimization algorithm module uses the predicted emission concentration and calculated combustion efficiency as targets, and employs the particle swarm optimization algorithm to solve for the optimized boiler operating parameters online.

[0039] S5. The optimized operating parameters are sent to the field gateway device via MQTT communication and written into the PLC controller to adjust the boiler operation.

[0040] S6. The system's operating status, data, and optimization results are updated and displayed on the screen in real time, while new data is stored in the database for model updates.

Claims

1. A detection and control system for a pulverized coal horizontal boiler, characterized in that, It includes a data acquisition module, a data processing module, a prediction model module, an optimization algorithm module, and a system execution and display module; The data acquisition module is used to collect boiler operating data in real time and transmit the operating data to the data processing module; The data processing module is used to preprocess the running data to obtain standardized time series data, and then transmit the standardized time series data to the prediction model module; The prediction model module has a built-in neural network model trained based on a long short-term memory network. It is used to receive the standardized time series data, output the predicted values ​​of nitrogen oxide and ammonia emission concentrations for future periods, and pass the predicted values ​​to the optimization algorithm module. The optimization algorithm module is used to calculate the optimized boiler operating parameters online using the predicted value as one of the optimization objectives, and then transmit the optimized operating parameters to the system execution module. The system execution and display module controls the boiler to operate according to the optimized operating parameters and displays the boiler's operating status in real time.

2. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The boiler's operating data includes gas flow rate, primary air volume, secondary air volume, flue gas oxygen content, furnace temperature, instantaneous coal feed rate, instantaneous ammonia flow rate, and NO content in the flue gas. x With NH3 concentration.

3. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The preprocessing described in the data processing module includes outlier removal, data format conversion, standardization, and serialization by time window.

4. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The feature extraction layer in the neural network structure of the prediction model module consists of two Conv1D layers and two MaxPooling1D layers. The time series processing layer consists of two LSTM layers; the output layer consists of two Dense layers.

5. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The fitness function constructed by the particle swarm optimization algorithm is the sum of the negative weighting of nitrogen oxide emission concentration and the positive weighting of combustion efficiency.

6. A detection and control system for a pulverized coal horizontal boiler according to claim 1 or 5, characterized in that, The optimized operating parameters calculated by the particle swarm optimization algorithm include the ammonia-to-coal ratio, primary air volume, and secondary damper opening.

7. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The real-time data displayed in the system execution and display module includes instantaneous ammonia flow rate, instantaneous pulverized coal flow rate, instantaneous air flow rate, combustion air oxygen content, and furnace temperature.

8. The detection and control system for a pulverized coal horizontal boiler according to claim 1, characterized in that, The system execution and display module also includes a database unit for unified storage and management of historical data, real-time data, model parameters, and user information.

9. A method of using a detection and control system for a pulverized coal horizontal boiler, for use with the detection and control system for a pulverized coal horizontal boiler as described in any one of claims 1 to 8, characterized in that, This includes the following steps performed sequentially: S1. Real-time acquisition of boiler operating data; S2. Preprocess the running data to obtain standardized time series data; S3. Input the standardized time series data into the pre-trained LSTM prediction model to obtain the future time period NO. x Compared with the predicted NH3 emission concentration; S4. The optimization algorithm module uses the predicted emission concentration and calculated combustion efficiency as targets, and employs the particle swarm optimization algorithm to solve for the optimized boiler operating parameters online. S5. The optimized operating parameters are sent to the field gateway device via MQTT communication and written into the PLC controller to adjust the boiler operation; S6. The system's operating status, data, and optimization results are updated and displayed on the screen in real time, while new data is stored in the database for model updates.

Images