论文部分内容阅读
矿物浮选过程中,矿浆pH值作为影响浮选效果的一个重要因素,是实现浮选过程监视及优化控制的一个重要参量。目前的pH值测定仪存在交叉污染、测量滞后等问题,难以获得实时准确的pH值。为使浮选运行在最优状态,在泡沫图像特征提取的基础上,提出一种基于自适应遗传混合神经网络的预测模型,该模型首先利用主元分析(PCA)方法对提取的多个图像特征进行降维,然后采用自适应遗传混合神经网络(AGA-HNN)建立pH值预测模型。最后将该模型应用于浮选现场,预测结果能够实时跟踪实际值,根据预测值实时调整工况条件,改善了浮选效果,提高了浮选效率。
Mineral flotation process, pulp pH value as an important factor affecting the flotation effect, is to achieve flotation process monitoring and optimization of control is an important parameter. The current pH meter has problems such as cross-contamination, measurement lag, and the like, making it difficult to obtain a real-time and accurate pH value. In order to optimize the flotation operation, a new prediction model based on adaptive genetic hybrid neural network is proposed based on the feature extraction of foam image. This model first uses PCA (principal component analysis) method to extract multiple images Feature reduction, and then use adaptive genetic mixed neural network (AGA-HNN) to establish a pH prediction model. Finally, the model is applied to the flotation site. The prediction result can track the actual value in real time and adjust the working conditions according to the predicted value in real time to improve the flotation effect and improve the flotation efficiency.