CN-121993800-A - Digital twin and AI feedforward-based self-adaptive accurate combustion control system and method for garbage incineration boiler

CN121993800ACN 121993800 ACN121993800 ACN 121993800ACN-121993800-A

Abstract

A self-adaptive precise combustion control system and a self-adaptive precise combustion control method for a garbage incinerator based on digital twin and AI feedforward belong to the technical field of incinerator combustion control. The problems of unstable combustion working condition, large temperature fluctuation range, unstable steam yield and parameters and poor control precision of the existing garbage incineration boiler power generation are solved, and the system comprises an online sensing system, a digital twin model, an AI feedforward-feedback cooperative controller and a boiler control system. The on-line sensing system detects the physicochemical characteristics of the garbage entering the furnace in real time, the digital twin model is constructed based on the first nature principle of combustion, the garbage characteristics and the real-time state of the boiler are combined, the combustion working condition of 3-5 minutes in the future is simulated and predicted, the AI controller generates an optimal feedforward control instruction through a pre-trained deep reinforcement learning model according to the prediction result, the fire grate speed, the air quantity and the auxiliary fuel are dynamically adjusted, and the boiler control system executes the instruction to realize accurate control. The invention is suitable for incinerator control.

Inventors

WU YANQIU
YE JIANQUN
ZHU HONGBAO
JIANG WENYAN
WU XIAOYU
ZHANG MINGBAO
SONG YONGFU
LU AIHUA
YANG HONGHONG
YUAN DEXIANG
LI YANDONG
DAI JIN
JIANG YUNHAO

Assignees

哈尔滨锅炉厂有限责任公司

Dates

Publication Date: 20260508
Application Date: 20260204

Claims (7)

1. The self-adaptive precise combustion control system of the garbage incineration boiler based on digital twinning and AI feedforward is characterized by comprising an online sensing system (1), a digital twinning model (2), an AI feedforward-feedback cooperative controller (3) and a boiler control system (4); The on-line sensing system (1) is configured on the garbage feeding device and is used for continuously detecting physical and chemical characteristic parameters of garbage before entering the incineration boiler in real time; The digital twin model (2) is a real-time dynamic simulation model constructed based on the principle of the first nature of the heat, flow and combustion of the garbage incineration boiler, physical and chemical characteristic parameters of the garbage and the current running state data of the boiler are used as the input of the real-time dynamic simulation model, and the combustion working condition parameters of the boiler in the next time period are obtained through simulation; The AI feedforward-feedback cooperative controller (3) is used for receiving the combustion working condition parameters of the boiler in the next time period, generating an optimal feedforward operation instruction for the next time period of the boiler executing mechanism by combining a pre-trained AI algorithm model, and transmitting the instruction to the boiler control system (4); The boiler control system (4) controls the boiler and auxiliary equipment to control parameters of the garbage incineration boiler by utilizing the instructions.
2. The self-adaptive accurate combustion control system of the garbage incineration boiler based on digital twin and AI feedforward as claimed in claim 1, wherein the online perception system (1) comprises a near infrared spectrometer, a microwave moisture meter and a visual recognition system; the near infrared spectrometer is arranged above or beside the conveying chain plate of the feeder and is used for qualitatively and quantitatively detecting key chemical components in the garbage by analyzing the position and the intensity of an absorption peak of the reflection or transmission spectrum of near infrared light on the surface of the garbage; The microwave moisture meter is used for non-contact collection of the overall volume moisture content of the garbage at the inlet of the garbage incineration boiler; The visual recognition system comprises an industrial camera and a light supplementing unit, is arranged above the feeding channel and is used for acquiring physical form information of the garbage by shooting images or video streams during garbage conveying.
3. The digital twin and AI feed forward based adaptive accurate combustion control system of a waste incineration boiler according to claim 1, further comprising a data storage and analysis platform for storing historical data from the on-line perception system (1), the digital twin model (2) and the controller (3).
4. The adaptive accurate combustion control system of a waste incineration boiler based on digital twinning and AI feedforward according to claim 1, characterized in that the digital twinning model (2) comprises a data perception and synchronization layer and a mechanism and physical model layer; the data sensing and synchronizing layer is used for acquiring real-time dynamic data constructed by a first principle of heating power, flow and combustion of the garbage incineration boiler; The mechanism and physical model layer utilizes the real-time dynamic data constructed by the first principles of heat, flow and combustion of the garbage incineration boiler to construct a constructed real-time dynamic simulation model, and utilizes the physical and chemical characteristic parameters of the garbage before entering the incineration boiler to simulate the reaction process of the garbage on the fire grate to acquire the combustion working condition parameters of the boiler in the next time period.
5. The adaptive precise combustion control system of the garbage incineration boiler based on digital twin and AI feedforward according to claim 4, wherein the AI feedforward-feedback cooperative controller (3) comprises a data-driven AI model layer, the data-driven AI model layer uses combustion working condition parameters of the boiler in the next time period, aims at maximum steam yield and boiler efficiency and minimum pollutant emission, coal consumption and corrosion rate under the constraint of an upper limit and a lower limit of the furnace temperature, an environment-friendly emission limit and an equipment safety limit, and obtains the speed of each stage of the grate at the next moment, the air quantity and the angle of twice air supply and an auxiliary fuel injection quantity control instruction as the optimal feedforward operation instruction in the next time period by using an AI algorithm model trained by a deep reinforcement learning algorithm.
6. A digital twinning and AI feedforward-based self-adaptive precise combustion control method for a waste incineration boiler, which is realized based on the system of any one of claims 1 to 5, and is characterized in that the method comprises the following steps: S1, acquiring physical and chemical characteristic parameters of garbage before entering an incineration boiler in real time; S2, inputting physical and chemical characteristic parameters of the garbage and current running state data of the boiler into a digital twin model (2), and obtaining combustion working condition parameters of the boiler in the next time period through simulation; S3, generating an optimal feedforward operation instruction based on a pre-trained AI algorithm according to the acquired combustion working condition parameters of the boiler in the next time period, and transmitting the instruction to a boiler control system (4), wherein the boiler control system (4) controls parameters of the garbage incineration boiler by controlling the boiler and auxiliary equipment.
7. The adaptive precise combustion control system of a waste incineration boiler based on digital twinning and AI feedforward as claimed in claim 6, wherein one of the combustion condition parameters of the boiler in the next time period is 3 to 5 minutes in future.

Description

Digital twin and AI feedforward-based self-adaptive accurate combustion control system and method for garbage incineration boiler Technical Field The invention belongs to the technical field of combustion control of an incinerator. Background The stable operation of the garbage incineration boiler is the key for ensuring the efficiency, the environmental protection standard and the service life of equipment. However, the prior art faces fundamental challenges: (1) The components, the moisture and the heat value of the household garbage entering the furnace are severely fluctuated on the time scale of the minute level, so that the combustion working condition in the furnace is unstable, and the phenomenon is that: and a, the temperature of the hearth fluctuates in a large range to influence the thorough decomposition of dioxin. B, unstable steam yield and parameters (pressure and temperature) affect the power generation quality and equipment safety. And c, air distribution is difficult to match accurately, so that insufficient combustion or excessive air coefficient is too high. (2) The hysteresis of the control system, namely the existing control system (DCS) mainly depends on feedback signals such as hearth temperature, oxygen amount, steam pressure and the like. When the sensor detects the parameter change and starts to adjust, the garbage working condition is changed, the control system is in a 'catch-up' working condition for a long time, and delay and overshoot exist in the adjustment. (3) And highly depends on manual experience, and when the working condition is changed drastically, manual intervention of an operator is still needed, so that real full-automatic and optimal operation is difficult to realize. Disclosure of Invention The invention aims to solve the problems of unstable combustion working conditions, large temperature fluctuation range, unstable steam yield and parameters and poor control precision of the existing garbage incineration boiler power generation, and provides a self-adaptive accurate combustion control system and method for the garbage incineration boiler based on digital twin and AI feedforward. The invention relates to a self-adaptive accurate combustion control system of a garbage incineration boiler based on digital twin and AI feedforward, which comprises the following components: the system comprises an online sensing system, a digital twin model, an AI feedforward-feedback cooperative controller and a boiler control system; the on-line sensing system is arranged on the garbage feeding device and is used for continuously detecting physical and chemical characteristic parameters of garbage before entering the incineration boiler in real time; The digital twin model is a real-time dynamic simulation model constructed based on the first principles of heat, flow and combustion of the garbage incineration boiler, physical and chemical characteristic parameters of the garbage and current running state data of the boiler are used as inputs of the real-time dynamic simulation model, and combustion working condition parameters of the boiler in the next time period are obtained through simulation; The AI feedforward-feedback cooperative controller is used for receiving the combustion working condition parameters of the boiler in the next time period, generating an optimal feedforward operation instruction for the next time period of the boiler executing mechanism by combining a pre-trained AI algorithm model, and transmitting the instruction to the boiler control system; and the boiler control system controls the boiler and auxiliary equipment to control parameters of the garbage incineration boiler by utilizing the instructions. Further, in the invention, the online sensing system comprises a near infrared spectrometer, a microwave moisture meter and a visual recognition system; the near infrared spectrometer is arranged above or beside the conveying chain plate of the feeder and is used for qualitatively and quantitatively detecting key chemical components in the garbage by analyzing the position and the intensity of an absorption peak of the reflection or transmission spectrum of near infrared light on the surface of the garbage; The microwave moisture meter is used for non-contact collection of the overall volume moisture content of the garbage at the inlet of the garbage incineration boiler; The visual recognition system comprises an industrial camera and a light supplementing unit, is arranged above the feeding channel and is used for acquiring physical form information of the garbage by shooting images or video streams during garbage conveying. Further, the invention further comprises a data storage and analysis platform, wherein the data storage and analysis platform is used for storing historical data from the online sensing system, the digital twin model and the AI feedforward-feedback cooperative controller. Further, in the invention, the digital twin model comprises a