Research On Soft-Sensor Modeling Of The Flotation Process Based On Multi-Source Information Fusion

Since the flotation process is a complex physical and chemical synthesis reaction process, it has the characteristics of strong nonlinearity and coupling. So the concentrate grade and flotation recovery is difficult to be obtained in real time. This paper proposes using the method which combined artificial neural network and soft sensor modeling predicts concentrate grade and flotation recovery. The specific works are as follows:In order to forecast the key technology indicators in the flotation process-concentrate grade, a feed-forward neural network soft-sensor modeling based on PSO-GSA algorithm is proposed. Although gravitational algorithm has better optimization capability, it has slow convergence and it is easy to fall into local optimum. This chapter optimizes the speed and position of gravitational algorithm to improve the convergence rate and prediction accuracy by using particle swarm optimization algorithm. Finally, using proposed algorithm optimizes the parameters of feed-forward neural network soft-sensor modeling, and predicts and simulates the key technology indicators of flotation process.Secondly, features extraction of flotation froth images and BP neural network soft-sensor model of concentrate grade optimized by shuffled cuckoo searching algorithm is proposed. The flotation froth images contain lot of information about the flotation process, so fourteen parameters including color, visual and shape in flotation froth images is extracted as input variables of soft-sensor model; Using isometric mapping method reduces the dimensions, in order to reduce the input dimension and network size of BP neural network. Finally, a shuffling cuckoo search algorithm of adaptive step optimized BP neural network soft sensor model is put forward and simulated.Finally, an echo state network soft-sensor model optimized by the improved glowworm swarm optimization algorithm in flotation process is proposed. Then the kernel principal component analysis(KPCA) method is used to reduce the dimensionality of the high-dimensional input vector composed by the flotation froth image characteristics and process datum and extract the nonlinear principal components in order to reduce the ESN dimension and network complex. The ESN soft-sensor model of flotation process is optimized by the GSO algorithm with congestion factor. And predict and simulate concentrate grade and flotation recovery.In conclusion, simulation results show that three kinds of neural network soft-sensing model can obtain good prediction effect, and enhance prediction accuracy of concentrate grade and flotation recovery and satisfy the real-time control requirements of flotation process.