Laboratory Study And Optimization Of Coal-blending Coking Without Fat Coal

Coal-blending coking was conducted and optimized on a bench scale in the absence of fat coal. Cold properties (crushing resistance strength M13 and abrasion resistance strength M3) and thermal properties (coke reactivity toward CO2 CRI and coke strength after reaction CSR) of coke can be improved and ash and sulfur contents can be reduced by means of blending coals and adjusting their proportions as well as adding leaning and bonding agents. Macroscopic structure of coke was studied by means of methods such as coal property analysis and coal coking performance analysis. Microscopic structure of coke was investigated by using x-ray diffraction. Results show, abrasion resistance strength is improved with suitable weight of breeze as leaning agent; Coal blend is upgraded with medium-temperature pitch as bonding which should be controlled less than 5%; As other proportions stay constant in a blend, there is a best value of gas coal content for optimum coke quality; Mineral in coal has a catalytic influence on coke reactivity toward CO2, and mineral in M02YL coal has the most significant effect; Micro crystalline structure is the most intrinsic factor influencing coke qualities. Crystalline structure parameters of coke show strong relationship with coalification or ratio of inert and active materials.Crucible coking and small oven coking were compared. Results show that the laboratory experiment feasibility is confirmed by oven test. The laboratory experimental coking method has really a certain practical value.BP neural network is chosen for coke quality prediction. The mapping relationship between properties of single coal and those of coke from coal blend was built up with three-level structure. Prediction errors for all parameters are within 1%, meaning accuracy are high and ntework adaptability is strong.