Training method and device of fly ash concentration prediction model and computer equipment

Abstract

The invention relates to a training method and device of a fly ash concentration prediction model and computer equipment. The method comprises the following steps: acquiring training sample data; inputting the historical fly ash concentration time sequence into a time sequence processing layer to obtain a fly ash concentration processing result, and inputting the key feature data into a feature data processing layer to obtain a key feature processing result: inputting the fly ash concentration processing result and the key feature processing result into a fusion layer to obtain a data fusion result; inputting a data fusion result into an output layer to obtain a predicted fly ash concentration time sequence, wherein the output layer comprises a prediction head configured according to an output target of the fly ash concentration prediction model; and training a time sequence processing layer, a feature data processing layer, a fusion layer and an output layer based on the predicted fly ash concentration time sequence and the label fly ash concentration time sequence to obtain a trained fly ash concentration prediction model. By adopting the method, the fly ash concentration can be effectively predicted, the fly ash concentration prediction model is optimized, and the prediction accuracy is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a training method, device and computer equipment for a fly ash concentration prediction model. Background technique [0002] At present, coal-fired power plants use dust removal equipment and use the maximum power of the dust removal equipment to absorb fly ash in the exhaust gas of the power plant. However, in practical applications, the emission of fly ash is not very high most of the time, and it is not necessary to use the maximum power of the dust removal equipment. [0003] In order to reduce the energy consumption of dust removal equipment, traditional physical models based on domain knowledge are usually used as predictive models to process and control fly ash accordingly. However, when establishing a traditional physical model based on domain knowledge, it is necessary to conduct detailed investigations on the operating modes of different power stations, and...

AI Technical Summary

Problems solved by technology

However, when establishing a traditional physical model based on domain knowledge, it is necessary to conduct detailed investigations on the operation modes of different power stations, and to analyze the relationship between a large number of relevant input physical data. The period of establishing the model is long and the cost is high, which is difficult to meet the needs of practical applications.

Images