The Research Of Soft Measurement Basedon Energy Demodulation And Neural Network Optimized By Genetic Algorithm

In industrial production, multiphase flow system is a very popular research direction, in which the prediction of flow pattern and the rate of flow under various flow patterns is an important research content. Because of the oil-gas-water three-phase flow information signal with characteristics difficult to predict, so the study in this direction is relatively hard. In this paper, by using energy demodulation algorithm for oil-gas-water three-phase flow signal feature extraction and neural network optimized by genetic algorithms we build a soft measurement model to accurately predict flow pattern and the rate of flow, which provides a feasible method.Firstly, according to that the oil-gas-water three-phase flow feature extraction is very difficult and the accuracy is not high, this paper focuses on the principle of energy demodulation algorithm, and applied the optimized algorithm to feature extraction of oil-gas-water three-phase flow. By using the algorithm we can extract four signals parameters to characterize the main features of oil-gas-water three-phase flow. In addition, the experimental results show that the algorithm has a good effect on feature extraction.Secondly, for forecasting the flow pattern and the rate of flow of the oil-gas-water three-phase, this paper construct a high-quality soft measurement model by neural network optimized by genetic algorithms. By using genetic algorithms to optimize the weights and thresholds of neural network algorithm, a soft measurement model can be achieved to improve the accuracy. Using the soft measurement model can greatly enhance the soft measurement prediction accuracy of the flow pattern and the rate of flow for the view of the performance.Finally, based on the oil signal monitoring system, a variety of feature extraction algorithms and oil-gas-water three-phase flow soft measurement model, we use energy demodulation method as the main feature extraction algorithm to choose feature vector to train the soft measurement model optimized by genetic algorithms and the results proved that the soft measurement model has a good performance. The soft measurement model is applied to the identification of flow pattern, and then we use the model to forecast oil-phase flow rate and gas-phase flow rate of oil-gas-water three-phase flow, and the experimental results show that the soft measurement model has good prediction ability.