Effective fault detection techniques can help flotation plant reduce reagents consumption,increase mineral recovery,and reduce labor intensity.Traditional,online fault detection methods during flotation processes have...Effective fault detection techniques can help flotation plant reduce reagents consumption,increase mineral recovery,and reduce labor intensity.Traditional,online fault detection methods during flotation processes have concentrated on extracting a specific froth feature for segmentation,like color,shape,size and texture,always leading to undesirable accuracy and efficiency since the same segmentation algorithm could not be applied to every case.In this work,a new integrated method based on convolution neural network(CNN)combined with transfer learning approach and support vector machine(SVM)is proposed to automatically recognize the flotation condition.To be more specific,CNN function as a trainable feature extractor to process the froth images and SVM is used as a recognizer to implement fault detection.As compared with the existed recognition methods,it turns out that the CNN-SVM model can automatically retrieve features from the raw froth images and perform fault detection with high accuracy.Hence,a CNN-SVM based,real-time flotation monitoring system is proposed for application in an antimony flotation plant in China.展开更多
Conventional feature description methods have large errors in froth features due to the fact that the image during the zinc flotation process of froth flotation is dynamic,and the existing image features rarely have t...Conventional feature description methods have large errors in froth features due to the fact that the image during the zinc flotation process of froth flotation is dynamic,and the existing image features rarely have time series information.Based on the conventional froth size distribution characteristics,this paper proposes a size trend core feature(STCF)considering the froth size distribution,i.e.,a feature centered on the time series of the froth size distribution.The core features of the trend are extracted,the inter-frame change factor and the inter-frame stability factor are given and two calculation methods of the feature factors are proposed.Meanwhile,the STCF feature algorithm was established based on the core features by adding the inter-frame change factor and the inter-frame stability factor.Finally,a flotation condition recognition model based on BP neural network was established.The experiments show that the recognition model has achieved excellent results,proving that the method proposed effectively overcomes the limitation of the lack of dynamic information in the existing traditional size distribution features and the introduction of the two factors can improve the classification accuracy to varying degrees.展开更多
Graphite is naturally floatable due to its hydrophobic pro pe rty and also soft and smears on other gangue particles, rendering the gangue mor e or less floatable too. Due to this reason it is important to concentrate...Graphite is naturally floatable due to its hydrophobic pro pe rty and also soft and smears on other gangue particles, rendering the gangue mor e or less floatable too. Due to this reason it is important to concentrate on ar eas such as suitable flotation reagents, depression agents, pH modifiers, and pa rticle size to be fed during the process. The paper surveys and analyses the sui table particle size to be fed to achieve high-grade concentrate. According to t h e work carried out the author suggested the ideal cost effective flotation f low sheet for improved results at Bogala Mines in Sri Lanka.展开更多
基金Projects(61621062,61563015)supported by the National Natural Science Foundation of ChinaProject(2016zzts056)supported by the Central South University Graduate Independent Exploration Innovation Program,China
文摘Effective fault detection techniques can help flotation plant reduce reagents consumption,increase mineral recovery,and reduce labor intensity.Traditional,online fault detection methods during flotation processes have concentrated on extracting a specific froth feature for segmentation,like color,shape,size and texture,always leading to undesirable accuracy and efficiency since the same segmentation algorithm could not be applied to every case.In this work,a new integrated method based on convolution neural network(CNN)combined with transfer learning approach and support vector machine(SVM)is proposed to automatically recognize the flotation condition.To be more specific,CNN function as a trainable feature extractor to process the froth images and SVM is used as a recognizer to implement fault detection.As compared with the existed recognition methods,it turns out that the CNN-SVM model can automatically retrieve features from the raw froth images and perform fault detection with high accuracy.Hence,a CNN-SVM based,real-time flotation monitoring system is proposed for application in an antimony flotation plant in China.
基金Project(U1701261)supported by the National Science Foundation of China,Guangdong Joint Fund of Key ProjectsProject(61771492)supported by the National Natural Science Foundation of ChinaProject(2018GK4016)supported by Hunan Province Strategic Emerging Industry Science and Technology Research and Major Science and Technology Achievement Transformation Project,China。
文摘Conventional feature description methods have large errors in froth features due to the fact that the image during the zinc flotation process of froth flotation is dynamic,and the existing image features rarely have time series information.Based on the conventional froth size distribution characteristics,this paper proposes a size trend core feature(STCF)considering the froth size distribution,i.e.,a feature centered on the time series of the froth size distribution.The core features of the trend are extracted,the inter-frame change factor and the inter-frame stability factor are given and two calculation methods of the feature factors are proposed.Meanwhile,the STCF feature algorithm was established based on the core features by adding the inter-frame change factor and the inter-frame stability factor.Finally,a flotation condition recognition model based on BP neural network was established.The experiments show that the recognition model has achieved excellent results,proving that the method proposed effectively overcomes the limitation of the lack of dynamic information in the existing traditional size distribution features and the introduction of the two factors can improve the classification accuracy to varying degrees.
文摘Graphite is naturally floatable due to its hydrophobic pro pe rty and also soft and smears on other gangue particles, rendering the gangue mor e or less floatable too. Due to this reason it is important to concentrate on ar eas such as suitable flotation reagents, depression agents, pH modifiers, and pa rticle size to be fed during the process. The paper surveys and analyses the sui table particle size to be fed to achieve high-grade concentrate. According to t h e work carried out the author suggested the ideal cost effective flotation f low sheet for improved results at Bogala Mines in Sri Lanka.
