System and method for optimizing energy consumption of dust removal fan of iron and steel plant

Abstract

The invention belongs to the technical field of steelmaking dust removal, and particularly relates to an energy consumption optimization system and method for dust removal fans in a steel plant, and the method comprises the following steps: S1, collecting operation data and energy consumption data of a plurality of dust removal fans through an intelligent electric meter and a PLC; s2, the visual AI + application server carries out frame cutting processing on the video stream; s3, calling an energy consumption optimization AI model module by the application server, and inputting preprocessed historical data and real-time data; s4, the intelligent regulation and control module pushes the set value to a PLC, and the rotating speed of the multiple dust removal fans is adjusted; s5, the model updating and version management module evaluates the performance of the model every 24 hours, triggers updating according to needs and records version information; and S6, the plant-level energy consumption large screen and the system / process-level energy consumption board display core indexes such as total electricity consumption and electricity saving quantity in real time. According to the system and method, the equipment operation state, energy consumption data and yield data are collected in real time, the equipment energy consumption dynamic monitoring and analysis function is provided, the control mode is intelligent, and electric energy waste is avoided.

Description

Technical Field

[0001] This invention belongs to the field of steelmaking dust removal technology, and particularly relates to an energy consumption optimization system and method for dust removal fans in steel plants. Background Technology

[0002] Steel plant production processes generate large amounts of smoke, dust, and other waste gases. These gases contain harmful substances such as sulfur dioxide and nitrogen oxides, posing significant risks to human health and the environment. Dust removal systems collect and treat these smoke and dust, effectively reducing their concentration in the air, protecting worker health, and minimizing environmental pollution.

[0003] The current dust removal system has the following prominent problems: (1) High energy consumption: Many wind turbine motors have a power of over 1000kW and operate at full or half load for a long time. The annual electricity consumption is huge and the electricity cost accounts for a high proportion. There is an urgent need for energy conservation and cost reduction. (2) Outdated control mode: Most fans adopt fixed frequency operation or "high speed / low speed" two-level switching mode, which cannot achieve stepless speed regulation, making it difficult to match the actual dust generation changes, resulting in excessive air volume during non-operation periods and serious waste of electricity; (3) Reliance on manual experience for adjustment: The dust removal system is still adjusted and controlled by the staff based on their experience. The response is slow and the adjustment is not timely, which affects the dust removal effect and cannot guarantee the optimal energy saving. Summary of the Invention

[0004] The purpose of this invention is to provide an energy consumption optimization system and method for dust removal fans in steel plants, which solves the problems of high energy consumption, outdated control modes, and reliance on manual experience in the current dust removal system operation.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A dust removal fan energy consumption optimization system for steel plants includes an equipment layer, an infrastructure layer, a data layer, a platform layer, an application layer, and a presentation layer, with each layer communicating and connecting in sequence. The equipment layer includes multiple dust collector fans, a PLC controller, a smart meter, and an AI bullet camera. Each of the dust collector fans is connected to the PLC controller and the smart meter, enabling the acquisition of operational data, energy consumption data, and dust image data. The infrastructure layer includes an ICG gateway and an industrial fiber optic switch. The PLC controller connects to the ICG gateway, which in turn connects to the industrial fiber optic switch and the AI ​​bullet camera. The data layer includes a visual AI+ application server, an application server, and an IoT / storage server. The industrial fiber optic switch connects the visual AI+ application server to the application server, and the visual AI+ application server connects to the application server. Data is transmitted through the ICG gateway and the industrial fiber optic switch. Both the visual AI+ application server and the application server are connected to the IoT / storage server. Data processing and storage are performed via a visual AI+ application server, an application server, and an IoT / storage server. The platform layer includes an energy consumption optimization AI model module and a model update and version management module. The IoT / storage server connects to the energy consumption optimization AI model module and the model update and version management module, providing support for modeling and iteration. The application layer includes an intelligent control module. The IoT / storage server connects to the intelligent control module. The energy consumption optimization AI model module, the model update and version management module, and the intelligent control module are interconnected, receiving model output parameters and pushing them to the PLC controller to achieve fan control. The display layer includes a plant-level energy consumption screen and a system / process-level energy consumption dashboard. The IoT / storage server connects to the plant-level energy consumption screen, and the plant-level energy consumption screen connects to the system / process-level energy consumption dashboard, displaying operating status and optimization effects.

[0006] Preferably, the application server uses a CPU4410Y (24 cores 2.0GHz), 64GB of memory, and deploys energy-optimized application services, AI model training and deployment services, relational databases, Redis in-memory databases, and time-series databases.

[0007] Preferably, the AI ​​bullet camera uses an 8-megapixel bullet camera with a fixed focal length mode, and is used for on-site environmental monitoring and detection.

[0008] Preferably, the intelligent control module supports automatic / manual mode switching, supports configuring dust removal tasks according to peak and off-peak electricity periods, and sets the cycle to 1-3 times / day or 10-15 times / week.

[0009] Based on the above system, a method for optimizing the energy consumption of dust removal fans in steel plants is proposed, with the following specific steps: S1: The smart meter and PLC controller collect the operating data and energy consumption data of multiple dust removal fans through the ICG gateway at a sampling interval of 1-5 seconds. The AI ​​bullet camera collects dust video streams and transmits them to the visual AI+ application server. S2: The visual AI+ application server performs 1-3 frame / s frame-slicing processing on the video stream, identifies smoke areas through deep learning object detection algorithms, analyzes dynamic changes by combining background subtraction, and calculates dust concentration based on image blur algorithm, with an error of ≤5%. S3: The application server calls the energy consumption optimization AI model module, inputs preprocessed historical data and real-time data, identifies the working conditions through the K-means clustering algorithm, and outputs the optimal frequency setting value in the 20-50Hz range through the particle swarm optimization algorithm, with an output delay of ≤10s. S4: The intelligent control module pushes the set value to the PLC controller to adjust the speed of multiple dust removal fans, and at the same time collects the data after control and feeds it back to the model; S5: The model update and version management module evaluates model performance every 24 hours, triggers updates as needed, and records version information. S6: The plant-level energy consumption screen and system / process-level energy consumption dashboard display key indicators such as total power consumption and power saving in real time.

[0010] Preferably, in step S3, the energy consumption optimization AI model module adopts a random forest regression model, and the hyperparameters are optimized by grid search. The validation set prediction accuracy is ≥95% and the mean square error is ≤0.02, ensuring generalization ability.

[0011] Preferably, in step S4, the intelligent control module pushes the set value every 3-5 minutes in automatic mode, and only outputs suggested values ​​in manual mode. Compared with the prior art, the beneficial effects achieved by the present invention are as follows: (1) The model of relevant parameters is established by visual AI+ application server for prediction and analysis. The generated model service needs to save energy and reduce consumption while ensuring the dust removal effect. (2) The dust removal fan is monitored and adjusted in real time through the application server to ensure that it reaches a state of high efficiency and low energy consumption, effectively reducing the energy consumption of the dust removal fan and improving energy utilization efficiency; (3) The operating parameters of the dust removal fan are controlled by the intelligent control module, the operation of the dust removal fan is automatically adjusted, stepless speed regulation is achieved, the actual dust generation changes are matched, and the errors and costs caused by human intervention are reduced; (4) The system and method collect equipment operating status, energy consumption data, production data, etc. in real time, provide dynamic monitoring and analysis functions for equipment energy consumption, assist in energy consumption analysis and decision-making, intelligent control mode, realize stepless speed regulation, match actual dust production changes, reduce air volume during non-operation periods, and avoid energy waste. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the structure in an embodiment of the present invention.

[0013] Explanation of reference numerals in the attached diagram: 1. Dust removal fan; 2. PLC controller; 3. Smart meter; 4. AI bullet camera; 5. ICG gateway; 6. Industrial fiber optic switch; 7. Visual AI+ application server; 8. Application server; 9. IoT / storage server; 10. Energy consumption optimization AI model module; 11. Model update and version management module; 12. Intelligent control module; 13. Plant-level energy consumption display screen; 14. System / process-level energy consumption dashboard. Detailed Implementation

[0014] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0015] like Figure 1 As shown, a dust removal fan energy consumption optimization system for steel plants includes an equipment layer, an infrastructure layer, a data layer, a platform layer, an application layer, and a presentation layer, with each layer communicating and connecting in sequence. The equipment layer includes six dust collector fans (1), a PLC controller (2), a smart meter (3), and an AI bullet camera (4). All six dust collector fans (1) are connected to the PLC controller (2) and the smart meter (3), enabling the acquisition of operational data, energy consumption data, and dust image data. The infrastructure layer includes an ICG gateway (5) and an industrial fiber optic switch (6). The PLC controller (2) connects to the ICG gateway (5), which in turn connects to the industrial fiber optic switch (6) and the AI ​​bullet camera (4). The data layer includes a visual AI+ application server (7), an application server (8), and an IoT / storage server (9). The industrial fiber optic switch (6) connects to the visual AI+ application servers (7 and 8), which are connected to each other. Data is transmitted through the ICG gateway (5) and the industrial fiber optic switch (6). Both the visual AI+ application servers (7 and 8) are connected to the IoT / storage server (9). The AI+ application server 7, application server 8, and IoT / storage server 9 perform data processing and storage. The platform layer includes an energy consumption optimization AI model module 10 and a model update and version management module 11. The IoT / storage server 9 connects to the energy consumption optimization AI model module 10 and the model update and version management module 11, providing support for modeling and iteration. The application layer includes an intelligent control module 12. The IoT / storage server 9 connects to the intelligent control module 12. The energy consumption optimization AI model module 10, the model update and version management module 11, and the intelligent control module 12 are interconnected, receiving model output parameters and pushing them to the PLC controller 2 to achieve fan control. The display layer includes a plant-level energy consumption screen 13 and a system / process-level energy consumption dashboard 14. The IoT / storage server 9 connects to the plant-level energy consumption screen 13, and the plant-level energy consumption screen 13 connects to the system / process-level energy consumption dashboard 14, displaying the operating status and optimization effects.

[0016] Based on the above system, a method for optimizing the energy consumption of dust removal fans in steel plants is proposed, with the following specific steps: S1: The smart meter 3 and PLC controller 2 collect the operating data and energy consumption data of 6 dust removal fans 1 through the ICG gateway 5 at a 5s sampling interval. The AI ​​bullet camera 4 collects the dust video stream and transmits it to the visual AI+ application server 7. S2: The visual AI+ application server performs frame-by-frame processing on the video stream (1 frame / s), identifies smoke areas using a deep learning object detection algorithm, analyzes dynamic changes using background subtraction, and calculates dust concentration based on an image blur algorithm with an error ≤5%. S3: Application server 8 calls energy consumption optimization AI model module 10. Energy consumption optimization AI model module 10 adopts random forest regression model, and optimizes hyperparameters through grid search. The prediction accuracy of the validation set is ≥95% and the mean square error is ≤0.02, ensuring generalization ability. Input the preprocessed historical data and real-time data of the past six months, identify the working conditions through K-means clustering algorithm, and output the optimal frequency setting value in the 20-50Hz range through particle swarm optimization algorithm, with an output delay of ≤10s. S4: The intelligent control module 12 pushes the set value to the PLC controller 2 to adjust the speed of the 6 dust removal fans 1. At the same time, it collects the data after control and feeds it back to the model. In automatic mode, the intelligent control module 12 pushes the set value once every 5 minutes. In manual mode, it only outputs the suggested value. S5: Model Update and Version Management Module 11 evaluates model performance every 24 hours, triggers updates as needed, and records version information; S6: The plant-level energy consumption screen 13 and the system / process-level energy consumption dashboard 14 display key indicators such as total power consumption and power saving in real time.

[0017] Before the energy consumption optimization system was put into use, the average current was 45.47A, and after it was put into use, the average current was 35.94A, a difference of 9.53A. The average current was significantly reduced before and after the system was put into use, and the dust concentration did not change significantly before and after the system was put into use.

[0018] Before and after the energy consumption optimization system was put into use, the power consumption of the double-flip dust collector fan was displayed as follows: the average daily power saving was 4416 kWh, the average power consumption of the dust collector fan was reduced by 5%, and the daily cost was reduced by 2164 yuan, while ensuring the dust removal effect.

Claims

1. An energy consumption optimization system for dust removal fans in steel plants, characterized in that, It includes the device layer, infrastructure layer, data layer, platform layer, application layer, and presentation layer, with each layer communicating and connecting in sequence. The equipment layer includes multiple dust collector fans, a PLC controller, a smart meter, and an AI bullet camera. Each of the dust collector fans is connected to the PLC controller and the smart meter, enabling the acquisition of operational data, energy consumption data, and dust image data. The infrastructure layer includes an ICG gateway and an industrial fiber optic switch. The PLC controller connects to the ICG gateway, which in turn connects to the industrial fiber optic switch and the AI ​​bullet camera. The data layer includes a visual AI+ application server, an application server, and an IoT / storage server. The industrial fiber optic switch connects the visual AI+ application server to the application server, and the visual AI+ application server connects to the application server. Data is transmitted through the ICG gateway and the industrial fiber optic switch. Both the visual AI+ application server and the application server are connected to the IoT / storage server. Data processing and storage are performed via a visual AI+ application server, an application server, and an IoT / storage server. The platform layer includes an energy consumption optimization AI model module and a model update and version management module. The IoT / storage server connects to the energy consumption optimization AI model module and the model update and version management module, providing support for modeling and iteration. The application layer includes an intelligent control module. The IoT / storage server connects to the intelligent control module. The energy consumption optimization AI model module, the model update and version management module, and the intelligent control module are interconnected, receiving model output parameters and pushing them to the PLC controller to achieve fan control. The display layer includes a plant-level energy consumption screen and a system / process-level energy consumption dashboard. The IoT / storage server connects to the plant-level energy consumption screen, and the plant-level energy consumption screen connects to the system / process-level energy consumption dashboard, displaying operating status and optimization effects.

2. The energy consumption optimization system for dust removal fans in steel plants according to claim 1, characterized in that, The application server uses a CPU4410Y (24 cores 2.0GHz) and 64GB of memory, and deploys energy-optimized application services, AI model training and deployment services, relational databases, Redis in-memory databases, and time-series databases.

3. The energy consumption optimization system for dust removal fans in steel plants according to claim 1, characterized in that, The AI ​​bullet camera uses an 8-megapixel bullet camera with a fixed focal length and is used for on-site environmental monitoring and detection.

4. The energy consumption optimization system for dust removal fans in steel plants according to claim 1, characterized in that, The intelligent control module supports automatic / manual mode switching and allows configuration of dust removal tasks according to peak and off-peak electricity periods, with a cycle of 1-3 times / day or 10-15 times / week.

5. A method for optimizing the energy consumption of dust removal fans in steel plants based on the system described in claim 1, characterized in that, The specific steps are as follows: S1: The smart meter and PLC controller collect the operating data and energy consumption data of multiple dust removal fans through the ICG gateway at a sampling interval of 1-5 seconds. The AI ​​bullet camera collects the dust video stream and transmits it to the visual AI+ application server. S2: The visual AI+ application server performs 1-3 frame / s frame-slicing processing on the video stream, identifies smoke areas through deep learning object detection algorithms, analyzes dynamic changes by combining background subtraction, and calculates dust concentration based on image blur algorithm, with an error of ≤5%. S3: The application server calls the energy consumption optimization AI model module, inputs preprocessed historical data and real-time data, identifies the working conditions through the K-means clustering algorithm, and outputs the optimal frequency setting value in the 20-50Hz range through the particle swarm optimization algorithm, with an output delay of ≤10s. S4: The intelligent control module pushes the set value to the PLC controller to adjust the speed of multiple dust removal fans, and at the same time collects the data after control and feeds it back to the model; S5: The model update and version management module evaluates model performance every 24 hours, triggers updates as needed, and records version information. S6: The plant-level energy consumption screen and system / process-level energy consumption dashboard display key indicators such as total power consumption and power saving in real time.

6. The method for optimizing energy consumption of dust removal fans in steel plants according to claim 5, characterized in that, In step S3, the energy consumption optimization AI model module adopts a random forest regression model, and the hyperparameters are optimized through grid search. The validation set prediction accuracy is ≥95% and the mean square error is ≤0.02, ensuring generalization ability.

7. The method for optimizing energy consumption of dust removal fans in steel plants according to claim 5, characterized in that, In step S4, the intelligent control module pushes the set value every 3-5 minutes in automatic mode, and only outputs suggested values ​​in manual mode.

Images