A multivariate molten iron quality soft sensor method based on robust random weight neural network

A neural network and soft sensor technology, applied in the biological neural network model, steel manufacturing process, computer simulation, etc., can solve the problems that cannot fully reflect the quality level of molten iron, many consumables, intermittent discontinuity, etc.

CN105608492BActive Publication Date: 2018-03-20NORTHEASTERN UNIV LIAONING
5 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Publication Date
2018-03-20

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

A multi-element molten iron quality soft measurement method based on a robust random weight neural network belongs to the field of blast furnace smelting automation control, especially a multi-element molten iron quality based on Cauchy distribution weighted M-estimation random weight neural networks (M‑RVFLNs). Parametric dynamic soft measurement method. This invention uses the principal component analysis (PCA) method to screen out the most important parameters that affect the quality of molten iron in the blast furnace as model input variables, and constructs a multivariate dynamic prediction model of molten iron quality that has an output self-feedback structure and takes into account input and output data at different times. It can simultaneously Multivariate dynamic soft measurement is performed on the main parameters Si content, P content, S content and molten iron temperature that characterize the quality of blast furnace hot metal. The invention includes the following steps: (1) auxiliary variable selection and model input variable determination; (2) training and use of M-RVFLNs soft measurement model.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the field of automatic control of blast furnace smelting, and in particular relates to a dynamic soft sensing method for multivariate molten iron quality parameters in the blast furnace ironmaking process based on Cauchy distribution weighted M-estimated random weight neural networks (M-RVFLNs). Background technique

[0002] Blast furnace ironmaking is used to reduce iron from iron ore and other iron-containing compounds to smelt qualified molten iron. The ironmaking process is an extremely complex nonlinear dynamic process. Blast furnace ironmaking reduces iron from iron ore and other iron-containing compounds through complex gas-solid, solid-solid, and solid-liquid reactions in the furnace, and smelts qualified molten iron. At the same time, the quality index of molten iron, as the most important production index in the blast furnace ironmaking process, directly determines the quality of subsequent steel products and the ener...

Examples

Embodiment Construction

[0087] As shown in the figure, the present invention is based on the computer system composition of conventional measurement system, data collector, M-RVFLNs soft sensor software and operating software, and the detailed structure is as follows figure 1 shown. Conventional measuring instruments such as flowmeters, pressure gauges and thermometers are installed in various corresponding positions of the blast furnace smelting system. The data collector is connected to the conventional measurement system, and connected to the computer system running the online forecast software through the communication bus. The conventional measuring system mainly includes the following conventional measuring instruments including:

[0088] Three flowmeters are used to measure the pulverized coal injection volume, oxygen-enriched flow, and cold air flow of the blast furnace pulverized coal injection system on-line;

[0089] A thermometer for online measurement of the hot blast temperature of th...