No abstract

An integrated fuzzy logic-neural network methodology is presented as a mean to improve the reconstruction of the performance map of axial compressors and fans. The learning capability of artificial neural network technique is integrated to the knowledge aspect of fuzzy inference system to offer enhanced prediction capabilities compared to using a single methodology independently. The proposed technique incorporates information of experimental data on surge, operating, and choke lines at any arbitrary but fixed rotational speed. A comparison of the predicted results with experimental data reveals a very good agreement. The proposed technique has the capability to model the nonlinear surge line as well as the kink in the performance map. Application of the method for compressor map generation showed that the proposed technique is robust and capable of enhancing any performance simulation tool used for the dynamic simulation and condition monitoring.

The application of artificial neural network to compressor performance map prediction is investigated. Different types of artificial neural network such as multilayer perceptron network, radial basis function network, general regression neural network, and a rotated general regression neural network proposed by the authors are considered. Two different models are utilized in simulating the performance map. The results indicate that while the rotated general regression neural network has the least mean error and best agreement to the experimental data, it is however limited to curve fitting application. On the other hand, if one considers a tool for curve fitting as well as for interpolation and extrapolation applications, multilayer perceptron network technique is the most powerful candidate. Further, the compressor efficiency based on the multilayer perceptron network technique is determined. Excellent agreement between the predictions and the experimental data is obtained.

In this article, a feed-forward neural network is explored to reconstruct the performance map of an axial compressor through the utilization of a limited number of experimental data. The Levenberg—Marquardt algorithm with Bayesian regularization method is used to adjust the weights and biases of the network. The proposed technique is utilized to estimate the mass flowrate, the pressure ratio, the shaft speed, and the efficiency in regions where no experimental data are available. The surge line is predicted and the line of maximum efficiencies is determined. The results are compared with experimental data.

A rotated general regression neural network is presented as an enhancement to the general regression neural network. A variable kernel estimate for multivariate densities is considered. A coordinate transformation is adopted which circumvent the difficulty of predicting multimodal distribution with large variance differences between modes which is associated with the general regression neural network. The proposed technique trains the network in a way that the variance differences between modes is kept small and in the same order. Further, the technique reduces the number of indispensable training parameters to two parameters and lowers the load of the computation as well as the time for conditions in which employing separate values of σ is unavoidable. The accuracy of the proposed technique is demonstrated by examining two different cases: the performance map of an axial compressor and the boundary layer profile over a flat plate. The results are compared with those by general regression neural network as well as the corresponding experimental data. Excellent improvement is obtained.