Intelligent monitoring regulation and control method and system in waste gas treatment system
By combining deep learning and reinforcement learning algorithms, an intelligent monitoring and control system was constructed, which solved the stability and energy consumption problems of traditional waste gas treatment systems under complex operating conditions, and realized real-time monitoring and efficient purification of waste gas pollutants.
Patent Information
- Application Number
- CN202511899409.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-16
- Publication Date
- 2026-03-17
AI Technical Summary
Traditional waste gas treatment systems struggle to cope with complex operating conditions, resulting in unstable control of pollutant emissions and high energy consumption.
A monitoring model is constructed using deep learning algorithms, and the equipment operating parameters are dynamically adjusted by combining reinforcement learning algorithms. Sensors are used for real-time monitoring and optimization strategies to achieve intelligent control.
It enables real-time monitoring and early warning of pollutant concentrations in exhaust gas, optimizes equipment operating parameters, reduces energy consumption, and improves pollutant removal rate and purification efficiency.
Smart Images
Figure CN121682122A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of waste gas treatment technology, and in particular to an intelligent monitoring and control method for a waste gas treatment system. Background Technology
[0002] In industrial production, the impact of waste gas emissions on the environment is becoming increasingly significant. For example, thermal power plants, steel smelters, and chemical plants generate waste gases containing large amounts of pollutants. These gases are complex in composition and their emission processes are often uncontrollable. Traditional waste gas treatment systems often rely on fixed operating parameters and manual adjustments, making it difficult to cope with complex changes in operating conditions and resulting in unstable control of pollutant emissions. To improve the efficiency and accuracy of waste gas treatment, artificial intelligence-based intelligent monitoring and control technologies have emerged. Summary of the Invention
[0003] In view of the above-mentioned problems, the present invention provides an intelligent monitoring and control method and system for waste gas treatment system. It uses deep learning algorithm to model the key parameters of the waste gas treatment system, predicts the concentration of abnormal waste gas pollutants, and provides early warning; and adopts optimization strategy to dynamically adjust the equipment operating parameters.
[0004] In a first aspect, the present invention provides an intelligent monitoring and control method for a waste gas treatment system, comprising: acquiring key parameter values in waste gas emissions collected in real time by sensor devices, and performing preprocessing operations to obtain data one; The data is used as input to the monitoring model to monitor the concentration of exhaust gas and several indicator data in real time. Construct a reinforcement learning model and learn strategies to dynamically adjust the operating parameters of the waste gas treatment equipment to ensure that the pollutant removal rate is not lower than the warning range; Real-time monitoring information, early warning information, and adjustment information of operating parameters of exhaust gas treatment equipment are transmitted to the user interface of the terminal control for display.
[0005] Furthermore, the sensor device includes an infrared sensor, an electrochemical sensor, a temperature / pressure sensor, and a flow / particulate matter sensor.
[0006] Furthermore, the monitoring model employs a time-series analysis algorithm, including a Conv-LSTM model, and the training steps of the monitoring model include: Data is collected from multiple sensor devices, then cleaned, normalized, and fused to obtain the data to be processed. Conv-LSTM framework is constructed, which has a convolutional neural network component to facilitate the extraction of data features, and then combined with LSTM to learn the temporal features of the data; The data to be processed is used as input to the Conv-LSTM framework. After multiple rounds of iterative training, a converged monitoring model is obtained.
[0007] Furthermore, the reinforcement learning model includes the PPO model, and its specific operation is as follows: Construct an action space with the operating parameters of each device as the operation object, and the action space adopts a network structure with several fully connected layers; The state space, which takes the data from each sensor as the object, is constructed using a multi-layer fully connected network structure. The reward space is designed with fan speed, spray volume, adsorbent dosage, and energy consumption as the reward and penalty factors. The reward function is as follows: ; Distributed model training is performed on the model constructed with action space, state space and reward function to obtain dynamic decision model.
[0008] Furthermore, the key parameter values include exhaust gas flow rate, temperature, particulate matter content, and pressure.
[0009] Secondly, an intelligent monitoring and control system for a waste gas treatment system includes the following modules: The data acquisition and processing module obtains key parameter values from the exhaust gas emissions collected in real time by the sensor devices, and performs preprocessing operations to obtain data one; The intelligent control module takes the data as input to the monitoring model, monitors the concentration of exhaust gas and several indicator data in real time; constructs a reinforcement learning model and learns strategies to dynamically adjust the operating parameters of the exhaust gas treatment equipment, ensuring that the pollutant removal rate is not lower than the warning range. The user interface display module transmits real-time monitoring information, early warning information, and adjustment information of the operating parameters of the exhaust gas treatment equipment to the user interface of the terminal control for display.
[0010] Thirdly, a computer-readable storage medium storing instructions that, when executed by a processor, implement the intelligent monitoring and control method in any of the above-mentioned waste gas treatment systems.
[0011] The above implementation scheme has the following advantages or beneficial effects: (1) Time series analysis algorithm is used to monitor the concentration of pollutants in exhaust gas in real time, which facilitates early warning; (2) Use reinforcement learning algorithms to optimize the strategy of dynamically adjusting equipment operating parameters, reduce energy consumption and improve pollutant removal rate; (3) Intelligent control technology is adopted to replace manual monitoring, improve the efficiency of monitoring and control, and improve the waste gas purification rate of the waste gas treatment system, thereby reducing energy consumption.
[0012] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0014] Figure 1 A flowchart of an intelligent monitoring and control method in a waste gas treatment system provided by the present invention; Figure 2 This invention provides a flowchart of an intelligent monitoring and control system in a waste gas treatment system. Detailed Implementation
[0015] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0016] Example 1 like Figure 1 The intelligent monitoring and control method shown in the waste gas treatment system includes the following steps: Step S01: Obtain the key parameter values of the exhaust gas emissions collected in real time by the sensor device, and perform preprocessing operations to obtain Data 1; Common sensor devices included in waste gas treatment systems include infrared sensors, electrochemical sensors, temperature / pressure sensors, flow / particulate matter sensors, and catalytic combustion sensors. During waste gas discharge, monitoring data from each sensor is collected, including waste gas flow rate, temperature, particulate matter content, and pressure. After real-time sampling, unreasonable values are removed, and default values are added. Then, data normalization is performed. For example, if the data ranges for temperature and waste gas flow rates differ significantly, directly using them as model input can easily lead to learning errors. Therefore, normalization is used to limit the values to a reasonable range. The resulting batch of sensor data is then stitched and fused to obtain the data to be predicted.
[0017] Step S02: Use the data as input to the monitoring model to monitor the concentration of exhaust gas and several indicator data in real time; In practical real-time applications, the detection model is trained using a large amount of data; the monitoring model employs time-series analysis algorithms, including Conv-LSTM and ARIMA models. The training steps of the monitoring model include: Data is collected from multiple sensor devices, then cleaned, normalized, and fused to obtain the data to be processed. Conv-LSTM framework is constructed, which has a convolutional neural network component to facilitate the extraction of data features, and then combined with LSTM to learn the temporal features of the data; The data to be processed is used as input to the Conv-LSTM framework. After multiple rounds of iterative training, a converged monitoring model is obtained.
[0018] The resulting monitoring model, after being lightweighted, is deployed to embedded and other hardware platforms for real-time forward inference in actual production.
[0019] The data to be tested obtained in the previous step is input into the deployed monitoring model, which uses real-time reasoning to predict whether there is a surge in exhaust pollutants or excessive exhaust flow that may prevent the decontamination reaction from being fully completed in the near future. Once a high-probability event of such a situation is predicted, corresponding countermeasures will be prepared to be taken at any time.
[0020] Step S03: Construct a reinforcement learning model and learn strategies to dynamically adjust the operating parameters of the waste gas treatment equipment to ensure that the pollutant removal rate is not lower than the warning range; Reinforcement learning models are used to learn a policy that maximizes benefit during interactions with the environment. Reinforcement learning models include DQN and PPO models. Specific operations are as follows: Construct an action space with the operating parameters of each device as the operation object, and the action space adopts a network structure with several fully connected layers; The state space, which takes the data from each sensor as the object, is constructed using a multi-layer fully connected network structure. The reward space is designed with fan speed, spray volume, adsorbent dosage, and energy consumption as the reward and penalty factors. The reward function is as follows: ; Distributed model training is performed on the model constructed with action space, state space and reward function to obtain dynamic decision model; the decision objective is to ensure that the pollutants detected by the sensor are removed and do not exceed the warning range.
[0021] The trained model is used in the above situations. When a high-probability event is predicted to occur, the adjustment strategy learned by the reinforcement learning model is used to analyze the current environmental conditions of the high-probability event and dynamically adjust the operating parameters of the waste gas treatment equipment so that the indicators of each sensor in the balanced environment do not exceed the warning range. This includes real-time data detected by various sensors. For example, the electrochemical sensor may detect that some pollutant chemical components still exceed the standard after treatment, but are within the warning range; the flow / particulate matter sensor may detect that the pollutant flow particles are too large, but are within the warning range.
[0022] Step S04: Transmit real-time monitoring information, early warning information, and adjustment information of operating parameters of the exhaust gas treatment equipment to the user interface of the terminal control for display.
[0023] The real-time monitoring information detected by each of the above sensors will be sent to the terminal and displayed on the terminal control interface in real time in the form of dynamic graphs. After a high-probability event is predicted, an early warning message will be sent to the terminal control terminal and displayed on the user interface to alert relevant personnel. The system sends all relevant parameters of the exhaust gas treatment equipment to the terminal, allowing staff to understand the equipment's operating status. Furthermore, after an alarm is triggered, the equipment automatically adjusts its relevant parameters and synchronizes them in real time to the user display interface on the control terminal.
[0024] Example 2 like Figure 2 As shown, an intelligent monitoring and control system in a waste gas treatment system includes the following modules: The data acquisition and processing module obtains key parameter values from the exhaust gas emissions collected in real time by the sensor devices, and performs preprocessing operations to obtain data one; The intelligent control module takes the data as input to the monitoring model, monitors the concentration of exhaust gas and several indicator data in real time; constructs a reinforcement learning model and learns strategies to dynamically adjust the operating parameters of the exhaust gas treatment equipment, ensuring that the pollutant removal rate is not lower than the warning range. The user interface display module transmits real-time monitoring information, early warning information, and adjustment information of the operating parameters of the exhaust gas treatment equipment to the user interface of the terminal control for display.
[0025] Example 3 A computer-readable storage medium storing instructions that, when executed by a processor, implement the intelligent monitoring and control method in any of the exhaust gas treatment systems described above.
[0026] The above solution achieves the following advantages or beneficial effects: (1) Time series analysis algorithm is used to monitor the concentration of pollutants in exhaust gas in real time, which facilitates early warning; (2) Use reinforcement learning algorithms to optimize the strategy of dynamically adjusting equipment operating parameters, reduce energy consumption and improve pollutant removal rate; (3) Intelligent control technology is adopted to replace manual monitoring, improve the efficiency of monitoring and control, and improve the waste gas purification rate of the waste gas treatment system, thereby reducing energy consumption.
[0027] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. An intelligent monitoring and control method in a waste gas treatment system, characterized in that, The method comprises the following steps: acquiring key parameter values in waste gas emission collected by a sensor device in real time, and performing a preprocessing operation to obtain data 1; using the data 1 as an input of a monitoring model to monitor the concentration of waste gas and several index data in real time; constructing a reinforcement learning model and learning a strategy of dynamically adjusting operation parameters of a waste gas treatment device to balance a pollutant removal rate not lower than a warning range; transmitting real-time monitoring information, warning information and adjustment information of the operation parameters of the waste gas treatment device to a user interface of terminal control for display.
2. The method of claim 1, wherein, The sensor device comprises an infrared sensor, an electrochemical sensor, a temperature / pressure sensor and a flow / particle matter sensor.
3. The method of claim 1, wherein the method further comprises: The monitoring model adopts a time series analysis algorithm, and a Conv-LSTM model is used; the training steps of the monitoring model comprise: collecting data of the multiple sensor devices, and then performing data cleaning, data normalization and data fusion and splicing to obtain data to be processed; constructing a Conv-LSTM framework having a convolutional neural network part to facilitate extraction of data features, and combining the LSTM to learn time sequence features of the data; using the data to be processed as an input of the Conv-LSTM framework, and performing multiple rounds of iterative training to obtain a converged monitoring model.
4. The method of claim 1, wherein the method further comprises: The reinforcement learning model comprises a PPO model, and the specific operation is as follows: constructing an action space taking operation parameters of each device as operation objects, and the action space adopts a network structure of several layers of full connection layers; constructing a state space taking sensor data as objects, and the state space adopts a network structure of several layers of full connection layers; The fan rotating speed, the spraying amount, the adsorbent feeding amount, and the energy consumption are taken as the reward space of the reward function, and the reward function is designed as: ; performing distributed model training on the model constructed by the action space, the state space and a reward function to obtain a dynamic decision model.
5. The method of claim 1, wherein the method further comprises: The key parameter values comprise waste gas flow, temperature, particle matter content and pressure value.
6. An intelligent monitoring and control system in a waste gas treatment system, characterized in that, The method comprises the following modules: a data acquisition and processing module for acquiring key parameter values in waste gas emission collected by a sensor device in real time, and performing a preprocessing operation to obtain data 1; an intelligent regulation and control module for using the data 1 as an input of a monitoring model to monitor the concentration of waste gas and several index data in real time, and constructing a reinforcement learning model and learning a strategy of dynamically adjusting operation parameters of a waste gas treatment device to balance a pollutant removal rate not lower than a warning range; a user interface display module for transmitting real-time monitoring information, warning information and adjustment information of the operation parameters of the waste gas treatment device to a user interface of terminal control for display.
7. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions are executed by the processor to implement the intelligent monitoring and regulation method in the waste gas treatment system according to any one of claims 1-6.