Method for predicating gas concentration in real time based on local decomposition-evolution neural network

Abstract

The invention relates to a method for predicating gas concentration in real time based on a local decomposition-evolution neural network. The method comprises the following steps that 1), the data of the gas concentration in a coal mine work face are acquired through a mine gas sensor, and the collected data of the gas concentration are stored in a historical database; 2), data in the gas concentration historical database are processed as a time sequence to obtain the data x (t) of the gas concentration time sequence, wherein time is the real time of collecting gas data, and the gas concentration is used as the dependent variable of the time; 3), LMD decomposition is carried out on the data x (t) of the gas concentration time sequence through a local decomposition algorithm to obtain a plurality of PF components; 4), neutral network modeling predication is respectively carried out on the obtained PF components; 5), the predication values of all the PF components are cumulated to obtain a gas emission amount predication result; 6), whether the gas monitor data of other monitor points need to be predicated or not is judged, if the gas monitor data of other monitor points need to be predicated, the step 3), the step 4) and the step 5) need to be repeated for predication, and if the gas monitor data of other monitor points do not need to be predicated, the predication is finished. The method can be widely applied to predicating the gas concentration in real time.

Description

technical field [0001] The invention relates to the field of gas concentration prediction in fully mechanized coal mining and tunneling working faces, in particular to a real-time gas concentration prediction method based on local decomposition-evolution neural network. Background technique [0002] Disasters caused by gas, as one of the main hidden dangers of coal mine safety production, have always been the focus of attention of relevant experts and scholars, and the gas disasters in coal mines in my country are particularly serious. For the monitoring and effective prediction of gas concentration in mine excavation and fully mechanized mining face, it can not only tap the potential ability of monitoring technology well, but also provide early warning for possible mine gas disasters, thereby reducing mine personnel and property losses. Then, the effectiveness and reliability of the technical support provided depends entirely on the accuracy and reliability of gas concentra...

AI Technical Summary

Problems solved by technology

[0003] In recent years, various prediction methods have been widely used in the prediction of gas concentration. A large number of previous empirical studies have shown that none of the existing prediction methods can well integrate the physical factors that affect the change of gas concentration, such as coal seam depth and coal seam thickness. , coal seam gas content, and coal seam spacing are taken into account, and the actual production conditions such as daily excavation and daily operation plan are not taken into account. There is a highly nonlinear relationship between these influencing factors, and the gas concentration prediction is only based on relevant experience.

Therefore, these traditional methods cannot effectively and accurately predict the gas concentration, and even some methods with artificial intelligence characteristics cannot establish an effective gas concentration prediction method

However, most coal mines can only provide historical data on gas concentration, and it is difficult to provide complete data on coal seam depth, coal seam thickness, coal seam gas content, coal seam spacing and other related factors. The difference has seriously affected the implementation effect of the gas concentration prediction results

Theoretically, this gas concentration prediction method has the following two defects: first, it relies too much on monitoring data, and does not analyze the synchronous behavior characteristics between them and real physical factors; Mine design has no guiding value

Images