The Aerodynamic Design Of Compact Axial Compressor And Research Of Geometric Uncertainties Quantification

At present,high-rotating compact axial compressors are widely used in the aviation,domestic appliance and unmanned aerial vehicle industries.However,there are few studies on the design method,flow field characteristics and noise of this kind of compressor published in the literature publicly.The design method,internal flow field characteristics and loss mechanism of the compact high speed axial compressor require further study compared with the traditional compressors.On the other hand,due to the small size,manufacturing uncertainty influence on its performance is very significant.In this paper,commercial software NREC is used to conduct the preliminary design of compact axial compressor,and according to the three dimensional calculation results,the corresponding geometrical adjustments of the compressor are adopted.The above process is repeated until its performance meets the requirements,finally end up with the design prototypes with single stage peak isotropic efficiency82.8% and flow rate along with pressure ratio of the desired goal of compressor whose overall size(diameter)is 26 mm.After preliminary performance meet the requirements,the product of the connect weights of neural network method is employed to carry out the sensitivity analysis of compressor’s rotor,the Back Propagation Neural Network(BPNN)is employed to construct the analysis relationship between the compressor rotor’s geometries and the performance based on the training and learning of the database,the prediction accuracy of the BPNN can be up to over 99.99%,which could exactly predict the compressor rotor’s performance.Then the modified Grason Algorithm based on the neural network connect weights is used to quantify the contribution of 22 compressor rotor geometrical effects on its performance.In order to eliminate the variation of single analysis,1000 analysis results is averaged and identified the key factors who mostly influence the compressor performance.The averaged analysis result shows that the tip clearance contributes 11.43%(efficiency sensitivity analysis)and 10.18%(pressure ratio sensitivity analysis)to compressor performance changes,while blade thickness distribution,front and trailing edge diameter and other single geometrical factors contribute about 4%,far less than the influence of tip clearance on compressor performance.This study focuses mainly on the robust optimization of compressor rotor tip clearance.Robust optimization involves the propagation of uncertainty in the system and uncertainty qualification.Non-intrusive probability collection point method(NIPRC)is adopted for this job.The robust optimization method based on BPNNsurrogate model coupled with multi-objective genetic algorithm NSGA II is employed to perform the optimization.Compared to the design prototype,the variance of robust compressor rotor’s efficiency could be reduced by 21.04%.Finally,10 cases of positively lean blades are designed,the CFD calculations are carried out and contrast analysis are performed.According to the analysis results,the compound lean blades are designed and the peak efficiency of the compressor rotor is increased by 1.09%,which is higher than that of the simple positively lean blades by 1.6%.In addition,the sensitivity analysis of the influence of the lean parameter of the blade on the compressor rotor performance was carried out in this paper based on the methods of variance decomposition analysis and the product of neural network connection weights.The analysis results show that the bending inclination is the most important factor affecting the compressor performance.