【作者】 边宝峰；

【导师】 马平；

【作者基本信息】 华北电力大学（河北） ， 控制理论与控制工程， 2005， 硕士

【摘要】 火电厂中钢球磨煤机筒内存煤量的测量问题一直是制粉控制效率低和自动控制难以投入运行的主要原因之一,针对D-S 证据理论存在的不足,而利用神经网络具有的自组织、自学习、并行分布处理、高度容错性和鲁棒性的特点,本文提出了一种将证据理论与神经网络相结合的数据融合方法,并将其用于解决球磨机存煤量的测量问题。融合结果表明该方法用于存煤量的测量能够有效判别出存煤量的数值范围及变化趋势,为球磨机自动控制的投入和运行操作提供了有效的保证。更多还原

【Abstract】 How to exactly measure the coal mass of pulverizer in fossil power plant is one of primary causes that the efficiency of milling control is low and the automation is difficult to running. Aiming at the limitation of evidential theory, As the neural network has self-organize, self-learn, collateral distributing manage, strong fault tolerant and robustness, we combined the evidential theory and the neural network and put forward a two-step fusion method. We can get the scale and the variety trend of the coal mass with this two-step fusion method. This would supply the effective guarantee to the automation and running of the ball milling plant.更多还原