Simulation Study on Oil Content Measurement of Two-Phase Flow Based on LSSVM Function
There is a strong interaction between two-phase fluids,with significant interface disturbances that can easily generate abnormal data points,affecting data collection and making it difficult to accurately reflect the specific heat characteristics of oil bodies,resulting in significant measurement errors in oil content.Therefore,the simulation of a two-phase flow oil content measurement method based on support vector machine is proposed.The sensing data is collected under the condition that the oil content of the two-phase flow system is known,and the singular spectrum a-nalysis method is introduced on the basis of the 3σ criterion to eliminate the abnormal data points caused by interfer-ence,so as to obtain effective data.The specific calorific value of the oil is calculated based on this to reflect its spe-cific heat characteristics.The LSSVM function is constructed by combining the least square method and the support vector machine algorithm.Through multiple training and learning,the generalization ability of the LSSVM function is improved to ensure measurement accuracy.Based on this,the obtained specific heat value of the oil body will be used as input,and the trained LSSVM function will be used to measure the oil content.Experiments show that after using the proposed method for measurement,the maximum NSE can reach 0.98,indicating that it can achieve accurate measurement of oil content in two-phase flow.
Singular spectrum analysisCalculation of specific heatLeast squares method
万方数据
维普
