Modeling method for boiler combustion optimization

Abstract

The invention relates to a modeling method of a boiler combustion optimization. The prior method can not solve the problem of boiler combustion optimization. In the invention, segmentation is carried out according to the active parameter load of combustion of a boiler, the load working conditions of similar combustion situations are segmented into one segment, the load working conditions with large differences are separately modeled, and a modeling method which is suitable for small samples and has high generalization ability is adopted for the low load working conditions with less data or extremely low load working conditions; selections which are distributed evenly and equal in quality on a topological structure are carried out for model data before modeling, a proper pretreatment is carried out to guarantee the predictive ability and the generalization ability of models, and finally the models corresponding to the load segments are selected to be optimized according to the load segment where a real load is positioned. The invention overcomes the defects that the traditional modeling method can not carry out modeling for all loads under the conditions that combustion situations are greatly different and guarantees the prediction accuracy and the generalization ability of the models through the data selection and the pretreatment.

Description

technical field [0001] The invention belongs to the technical field of information control and relates to pattern recognition and regression technology, in particular to a modeling method for boiler combustion optimization. Background technique [0002] Boiler combustion optimization is an important technical means for energy saving and emission reduction. Its goal is to obtain high efficiency, low pollution emissions and safer operating conditions by adjusting boiler air distribution, coal feeding and other operating parameters under certain boiler load conditions. The operation of the boiler is based on the load as the active parameter, and other operating parameters must be adjusted accordingly according to the boiler load. The configuration of operating parameters such as air distribution, coal supply and oxygen amount will directly lead to different boiler efficiencies, pollutant gas emissions and equipment loss. For a given boiler, under certain load conditions, there...

AI Technical Summary

Problems solved by technology

There is a very complex coupling relationship between the operating parameters of the boiler, and it is very difficult to find the optimal operating parameter configuration. In actual operation, due to the need for peak regulation, sometimes the boiler operates under low or extremely low load conditions. The operating conditions under low load or extremely low load conditions are very different from normal load and high load conditions. The combination of operating parameters such as air distribution and coal supply also has a great difference on the combustion state of the boiler. The actual operation of the boiler There are relatively few cases of medium and very low loads, and the amount of data will be relatively small. Therefore, trying to establish a single model for all loads through data mining to achieve optimal operation of the boiler usually encounters many difficulties in model prediction.

With the continuous advancement of science and technology, the degree of automation of boiler operation has been continuously improved, but the problem of boiler combustion optimization has not been well resolved.

[0003] In actual production, the combustion optimization of the boiler mainly depends on the commissioning personnel to conduct experiments under different working conditions. According to the specific boiler and coal type, different experiments are used to find a better operating parameter configuration. This method is time-consuming, laborious and can be tested. The parameter combination is limited, so there is still a lot of room for improvement in the parameter configuration found through debugging experiments. In addition, this method cannot realize online optimization according to the real-time changes of the boiler.

Images