Automatic Control System Of Dust Fan Based On Graphics Analysis

A lot of smoke is produced during the frequency melting process. To avoid polluting the environment, the smoke is drawn out through the pipes on the furnace. Economically, the speed of the induced draft fan (IDF) (the amount of smoke which is drawn out) should vary according to the concentration of smoke, which is high with the lots of smoke and low with little smoke. In order to measuring the concentration of the dust in the frequency melting furnace, this thesis develops a soft sensor system based on graphics analysis, wherein the smoke concentration in the furnace is measured indirectly by graphical features. The rotational speed of the An induced draft fan (IDF) is then adjusted in real time depending on the smoke concentration levels, which is classified by the distribution of the data of graphic features. The effect of the system has been remarkable since putting into production.The content of this thesis in brief includes:1. Through the investigation, the frequency melting process is divided into several operation stages. The required graphical features are selected after an analysis over the smoke condition in each specific stage.2. However, the brightness in the furnace is severely interfered by the condition of illumination in practice. To eliminate such interference, a compensation adjustment is proposed to get data in better quality for further analysis.3. The smoke conditions and screened graphs are quite distinguishing in different operation stages, while similar in the same stage. The rotational speed of the IDF can be set manually for few typical scenarios, but not able to do so in most cases. Thus, a semi-supervised clustering method is proposed to find the relations in data for setting an acceptable IDF speed.4. An automatic control system is developed. It is put into operation and works well.