|
Anionic reverse flotation process flotation recovery has been a very important economic indicators , it is the most important Peugeot concentrator technology and management level , so the need to accurately predict on the recovery , in order to better control . However , due to between flotation recovery its impact factors have strongly nonlinear uncertainties , difficult to use a precise mathematical model to describe the characteristics of a long time , has been restricting the establishment of the flotation recoveries measurement model . In order to solve the above problems , this paper , based on the actual production of a concentrator flotation , put forward a new recovery prediction method . The program uses the improved weighted LS-SVM construct predictive models to the froth image features as model input , LS-SVM parameters optimization through genetic algorithms , to make up for the past, a lot of inadequacies . The article first collected on-site foam image feature extraction , dig out the red component , certain correspondence between the bubble size and bubble speed , carrying capacity , crushing rate value flotation recoveries . And data classification using fuzzy C- means clustering method , using genetic neural network algorithm for data cleansing eliminate some undesirable data , clear data contains noise , incomplete and inconsistent data . Then further study the relationship between the eigenvalues ??and flotation recovery of the bubble image analysis algorithm inadequacies , this paper uses an improved LS-SVM foam image features to flotation recovery prediction model , and Matlab simulation platform test . The results show that the model established by using improved LS-SVM can more accurately predict the flotation recovery indicators , in line with the production process needs . Finally, flotation recovery prediction software design , the query prediction function and recovery data visualization . The software module development is based on the the VB.NET platform and Matlab platform to achieve . VB.NET mainly to complete the interface design and programming algorithm , Matlab platform to complete the online trend graph algorithm programming design , and for the VB.NET call , complementary advantages .
|