CN-122018581-A - Method for improving adding precision of out-of-stock agent and controlling response speed of system

Abstract

The invention relates to a method for improving the adding precision of a denitration agent and controlling the response speed of a system, and belongs to the technical field of control methods of sintering desulfurization and denitration systems. The technical scheme includes that a sensor system is used for acquiring system running state parameters, a data training set is prepared, a design model of the consumption of a stripping agent is built, the data training set is used for learning, model control parameters are obtained through online running, a cascade PID regulator taking the concentration of NOx at a chimney outlet as a set value and the frequency of an ammonia water motor as an output value is built in an L1 system, and the model control parameters are used as feedforward of a controller and used for L1 closed-loop control to adjust the running frequency of the ammonia water motor. The invention has the advantages that the control precision of the addition of the out-of-stock agent and the response speed of the actuator can be obviously improved, the invention is applicable to various industrial application scenes, and the control effect can be continuously optimized through a machine learning model.

Inventors

• YU TAO
• WU CHEN
• WANG YAN
• ZHANG CHI
• LI BAOZHONG
• LIU JIANPENG
• DONG HONGWANG
• LIU YAN
• ZHANG DA
• LI WENFENG
• LIU XIANG

Assignees

• 唐钢国际工程技术有限公司
• 唐山钢铁集团有限责任公司
• 河钢乐亭钢铁有限公司

Dates

Publication Date

20260512

Application Date

20260228

Claims (4)

1. 1. The method for improving the adding precision of the pin removing agent and controlling the response speed of the system is characterized by comprising the following steps: S1, data are collected in advance, system running state parameters are obtained through a sensor system, and a data training set is prepared; s2, constructing a model for predicting the consumption of the anti-sales agent, wherein the model for predicting the consumption of the anti-sales agent is obtained based on machine learning training of a neural network algorithm and is used for initializing model control parameters; s3, model learning and parameter generation, namely after the running state parameters of the system in the step S1 are processed, feeding the processed running state parameters into a dissales agent consumption prediction model in the step S2 for learning; S4, constructing a regulator, namely constructing a cascade PID regulator taking the concentration of NOx at a chimney outlet as a set value and the frequency of an ammonia water motor as an output value in an L1 system; S5, taking the recommended value of the addition amount of the out-of-stock agent as a feedforward parameter of the cascade PID regulator, taking the recommended value as the inter-stage reference calculation data of the regulator to participate in L1 closed-loop control, and adjusting the operation parameters of the ammonia water motor so as to realize high-precision control based on model prediction.
2. 2. The method for improving the adding precision of the denitration agent and controlling the response speed of the system according to claim 1, wherein in the step S1, the collected system operation state parameters comprise the amount of flue gas at the outlet of the dust remover, the amount of NO 2 at the outlet of the dust remover, the temperature at the outlet of the direct-fired furnace, the pressure of flue gas at the outlet of the direct-fired furnace, the amount of NO at the outlet of the chimney, the amount of NO 2 at the outlet of the chimney and the NH 3 content at the outlet of the chimney.
3. 3. The method for improving the dosing accuracy of a stripping agent and controlling the response speed of a system according to claim 1, wherein in the step S1, each system operation state parameter collecting position is located before the ammonia reactor, and a pre-calculation is provided for the PID setting parameters, and the calculated values form a real system feedforward.
4. 4. The method for improving the adding precision of the out-of-stock agent and the response speed of the control system according to claim 1, wherein in the step S4, the construction steps of the cascade PID regulator are as follows: s4.1 construction of the Pre-regulator The front-stage regulator is a nitrogen oxide regulator, the set value is the expected content of nitrogen oxide, the unit is mg/Nm 3 , the feedback value is the actual feedback value of nitrogen oxide, the unit is the same as the set value, and a proportional integral regulator is adopted; s4.2 construction of the post-regulator The rear-stage regulator is a flow regulator of the stripping agent, the set value is a weighted calculation value of the output and model control parameters of the front-stage regulator, the unit is kg/h, the feedback value is an actual feedback value of the stripping agent, the unit is the same as the set value, the output value of the regulator is a speed set value of a stripping agent adding device, the unit is rpm, and the proportional integral regulator is adopted.

Description

Method for improving adding precision of out-of-stock agent and controlling response speed of system Technical Field The invention relates to a method for improving the adding precision of a denitration agent and controlling the response speed of a system, and belongs to the technical field of control methods of sintering desulfurization and denitration systems. Background In the field of industrial automation, high-precision control of equipment is a key to improving production efficiency and reducing consumption. In the traditional sintering desulfurization and denitrification production process, ammonia water is added in advance, after chemical reaction of a reactor, the ammonia water is conveyed to the tail end of a chimney through a pipeline to be detected to obtain nitrogen oxide content feedback so as to regulate and control the addition amount of a denitration agent, and meanwhile, the nitrogen oxide content at the outlet of the chimney is closely related to various factors such as temperature and pressure in production, catalyst activity and the like. The traditional adding control method of the out-of-stock agent is generally based on a fixed control rule, such as simple PID (proportion-integral-derivative) process control, and is difficult to adapt to complex and variable working condition environments with multiple variables and hysteresis, so that the defects of low control precision and response speed, high production cost and the like exist. Disclosure of Invention The invention aims to provide a method for improving the adding precision of a pin removing agent and the response speed of a control system, which can obviously improve the controlling precision of the pin removing agent adding and the response speed of an actuator by adopting big data field extraction meeting characteristic values, feedforward signal position selection and synthesis, neural network modeling and basic automatic control technology, is suitable for various industrial application scenes, can continuously optimize the control effect through a machine learning model and effectively solves the problems in the prior art. The technical scheme of the invention is that the method for improving the adding precision of the out-of-stock agent and controlling the response speed of the system comprises the following steps: S1, data are collected in advance, system running state parameters are obtained through a sensor system, and a data training set is prepared; s2, constructing a model for predicting the consumption of the anti-sales agent, wherein the model for predicting the consumption of the anti-sales agent is obtained based on machine learning training of a neural network algorithm and is used for initializing model control parameters; s3, model learning and parameter generation, namely after the running state parameters of the system in the step S1 are processed, feeding the processed running state parameters into a dissales agent consumption prediction model in the step S2 for learning; S4, constructing a regulator, namely constructing a cascade PID regulator taking the concentration of NOx at a chimney outlet as a set value and adding ammonia water into an L1 system, wherein the frequency of the ammonia water is an output value; S5, taking the recommended value of the addition amount of the out-of-stock agent as a feedforward parameter of the cascade PID regulator, taking the recommended value as the inter-stage reference calculation data of the regulator to participate in L1 closed-loop control, and adjusting the operation parameters of the ammonia water motor so as to realize high-precision control based on model prediction. In the step S1, the collected system operation state parameters include the dust collector outlet flue gas amount, the dust collector outlet NO content, the dust collector outlet NO 2 content, the direct-fired furnace outlet temperature, the direct-fired furnace outlet flue gas pressure, the chimney outlet flue gas NO content, the chimney outlet flue gas NO 2 content and the chimney outlet flue gas NH 3 content. In the step S1, each system operation state parameter acquisition position is located before the ammonia reactor, and provides a pre-calculation for the PID setting parameters, and the calculated values form a real system feedforward. In the step S4, the construction steps of the cascade PID regulator are as follows: s4.1 construction of the Pre-regulator The front-stage regulator is a nitrogen oxide regulator, the set value is the expected content of nitrogen oxide, the unit is mg/Nm 3, the feedback value is the actual feedback value of nitrogen oxide, the unit is the same as the set value, and a proportional integral regulator is adopted; s4.2 construction of the post-regulator The rear-stage regulator is a flow regulator of the stripping agent, the set value is a weighted calculation value of the output and model control parameters of the front-stage regulator, the unit is kg/h, the fe