Compressor Performance Prediction Based on the Interpolation Method and Support Vector Machine

Compressors are important components in various power systems the field of energy and power. In practical applications, compressors often operate under non-design conditions. Therefore, accurate calculation on performance operating conditions is great significance for development application certain equipped with compressors. To calculate predict a compressor all through limited data, interpolation method was combined support vector machine (SVM). Based known data points design conditions, adopted to obtain training samples SVM. process, preliminary screening conducted kernel functions Two methods, including linear cubic spline interpolation, were used sample data. subsequent process SVM, genetic algorithm (GA) optimize its parameters. After training, available compared predicted The results show that SVM uses Gaussian function achieve highest prediction accuracy. accuracy trained obtained from higher than interpolation. Compared back propagation neural network optimized by (GA-BPNN), optimization extreme learning (GA-ELMNN), generalized regression (GA-GRNN), (GA-SVM) has better generalization, GA-SVM more predicting boundary GA-BPNN. addition, reducing number original still enables maintain high level predictive