A machine learning-based boiler fuel predictive control method and system

Abstract

The invention discloses a boiler fuel prediction control method and system based on machine learning, relates to the technical field of machine self-learning, and aims to solve the problem of poor control performance of the traditional temperature cascade PID method for boiler fuel prediction control. The main point of the technical solution is that the method includes: obtaining the temperature signal after water spraying; judging whether the temperature signal after water spraying is normal; learning the corrected upper limit assignment of the temperature target after water spraying according to the real-time temperature value after water spraying and the preset cumulative learning algorithm OT and the temperature target lower limit assignment OB after water spraying; the temperature target upper limit assignment OT after water spraying obtained through learning and the temperature target lower limit assignment OB after water spraying limit the main output of the superheated main steam temperature to the sprayer of the secondary temperature adjustment after the superheated water spraying The temperature target value range after water. The invention can maintain the stability of the main steam temperature without manual intervention, and has the effect of effectively improving the control performance.

Description

technical field [0001] The invention relates to the technical field of machine self-learning, in particular to a machine learning-based boiler fuel predictive control method and system. Background technique [0002] Both thermal power and thermal power use boilers for energy conversion, turning water into high-temperature and high-pressure steam. The quality of steam is measured by temperature and pressure indicators. Among them, the steam temperature is an important indicator. The temperature should not be too low, otherwise the superheat of the steam will be insufficient, and the water vapor will impact the final stage blade of the turbine, which will bring equipment orders and safety hazards; the temperature should not be too high, otherwise, the boiler overheating will easily cause the steam pipeline Metal fatigue and creep occur under high temperature and pressure, resulting in reduced pipe life. Therefore, in industrial sites with boiler containers, temperature contro...

AI Technical Summary

Problems solved by technology

[0005] However, the above-mentioned existing technical solutions have the following defects: the temperature target after water spraying is set according to the mathematical model of the steam temperature adjustment object, and the time-varying, uncertain and nonlinear nature of the steam temperature adjustment object makes it difficult to Establishing an accurate mathematical model and only relying on PID control, no matter how the PID parameters are matched, it is difficult to adapt the steam temperature to changes in various disturbances (such as changes in combustion conditions, dust accumulation in superheaters, large load fluctuations, soot blowing, etc.)

Moreover, once the operating conditions change greatly, the dynamic characteristics and model parameters of the steam temperature regulation object will be greatly affected, resulting in large temperature fluctuations

Therefore, it is difficult to obtain satisfactory control performance by using the traditional PID control method

Images