【摘要】 神华大柳塔煤和兖州北宿煤都是气流床气化的优良煤种.通过三种算法的比较,采用BP神经网络的L-M算法,分析煤的制焦终温与制焦升温速率、气化反应温度、Vdaf和煤焦H/C原子比等不同因子对煤焦气化反应速率模型预测精度的影响,建立了基于Matlab下神华大柳塔单煤种四因子和神华-兖州双煤种五因子煤焦高温气化反应速率神经网络预测模型,得到比较满意的结果,其相对误差分别是0.167和0.264.更多还原

【Abstract】 In this paper, two typical coals for entrained flow gasifiers,Shenhua coal and Yanzhou coal were used for raw materials. The L-M algorithm was employed in the BP neuron network through the comparison of three different kinds of algorithms, and the effects of factors, such as the content of volatile matter(Vdaf) in raw coals, the ratio of H and C (H/C) in coal chars, pyrolysis temperature, heating rate and gasification temperature, on the prediction error of BP neuron network were investigated. The neuron network with four factors were applicable for predicting the gasification rates of a single coal (Shenhua coal), while the neuron network with five factors for predicting those of different coals (Shenhua coal and Yanzhou coal). The result showed that the corresponding relative errors, which were respectively 0.167 and 0.264, were quite little.更多还原