Research And Application Of Sintering SO₂ Emission Model In Iron And Steel Enterprises

In thirteenth five-year planning,accelerating the improvement of environmental quality rise to the strategic position.In the planning,it is pointed out that the total amount of chemical oxygen demand,sulfur dioxide?SO₂?,ammonia nitrogen and nitrogen oxides should be controlled.Sintering is an important part of steel production and the main source of SO₂ emissions.SO₂ emissions generated for iron and steel production?excluding power plant?above 70%of the annual emissions in the sintering process.So the establishment of sintering flue gas SO₂ emission prediction model has an important guiding significance for achieving online prediction of SO₂emission and controling SO₂ emission.?1?Based on field investigation of sintering production line in a steel plant,the material flow,the energy flow and the sulfur flow of sintering unit were deeply analyzed.The model of material flow and sulfur flow balance was established.The distribution coefficient method was adopted.And then the calculation formula of material flow and sulfur flow distribution coefficient was derived.?2?Prediction model of SO₂ emission from sintering flue gas in a steel plant was established.The correlation between SO₂ emission of sintering flue gas and sulfur bearing materials was analyzed.A linear fitting model was established.And then,the positive and reverse prediction and the error calculation of sintering flue gas SO₂emissions were obtained by using the model.The results show that the maximum relative error of sulfur content in the forward prediction of sintering flue gas was 30%and the minimum value was 0.03%,and the average error was 13%;the maximum relative error of sulfur content in the reverse prediction of blended coke was 50%,the minimum was 0.07%,and the average error was 26%.In order to improve the error,the nonlinear fitting was carried out.Although the error has been improved,it is still larger.?3?The measured SO₂ concentration data of sintering flue gas?flue gas sulfur?conducted by the sliding average method and the wavelet threshold noise elimination method.Then the positive prediction model of BP neural network was established.The average mean square error?MSE?is 0.97%.The prediction results of the model were compared with the actual values,which can obtain that the average error was3.7%,and the maximum relative error was 5%,and the minimum relative error was0.1%,and the correlation was 0.96.According to the law of prediction model,the flue gas sulfur content increased by the blended ore with coke powder sulfur content increasing.?4?Based on visual studio design platform,the online prediction system of sintering flue gas SO₂ was developed by C#programming language.The system function includes 5 aspects:on-line prediction of S elements in sintering unit,SO₂prediction and warning,alarm,evaluation,query and report forms.