Soft measurement method for floating fine coal ash from slurry based on data drive

Abstract

The invention discloses a soft measurement method of a coal slurry flotation cleans ash based on date-driven, which establishes a hardware platform and a flotation cleans ash. Flotation feeding concentration, flow, flotation medicament dosage and tailings flashboard position are used as auxiliary variables, the flotation cleans ash is used as a master variable, the auxiliary variables are used as input and the flotation cleans ash is used as model output so as to carry out soft measurement modeling, and calibration and parameter correction are carried out to a model so as to estimate the ash content of the current flotation cleans in real-time. The method has simple application and low cost, can realize the soft measurement of the flotation cleans ash by only needing a small quantity of samples, thus improving the measurement timelessness, and being beneficial to realizing the optimized control of the flotation procedure.

Description

technical field [0001] The invention relates to the field of soft measurement technology in the field of industrial process control, in particular to a soft measurement method for the ash content of flotation clean coal in the coal slime flotation process. Background technique [0002] Soft measurement is an effective way to solve the situation that some key parameters need to be measured urgently but there is no corresponding test instrument, or the test instrument is too expensive. Specifically, soft measurement is to use the variables that need to be concerned about (that is, the variables to be measured) as the leading variables, and the related variables (measurable variables) that cause changes in the leading variables to be used as auxiliary variables, and the data of the auxiliary variables are used as input. The method of outputting the data and building a mathematical model to obtain the value of the leading variable, so soft measurement is sometimes called soft in...

AI Technical Summary

Problems solved by technology

[0004] The problem with this control strategy is that it is impossible to know the ash content of the flotation clean coal product in real time, so that the quality of the flotation clean coal is stuck.

Traditional statistical methods study the asymptotic theory when the number of samples approaches infinity, but it is difficult to achieve ideal results when the number of samples is limited or when dealing with small samples

Among them, the neural network has problems such as slow convergence speed, structure selection problem and local minimum, over-learning and poor promotion ability.

Images