Surface Defect Detection and Evaluation Method of Large Wind Turbine Blades Based on an Improved Deeplabv3+Deep Learning Model

The accumulation of defects on wind turbine blade surfaces can lead to irreversible damage,impacting the aero-dynamic performance of the blades.To address the challenge of detecting and quantifying surface defects on wind turbine blades,a blade surface defect detection and quantification method based on an improved Deeplabv3+deep learning model is proposed.Firstly,an improved method for wind turbine blade surface defect detection,utilizing Mobilenetv2 as the backbone feature extraction network,is proposed based on an original Deeplabv3+deep learning model to address the issue of limited robustness.Secondly,through integrating the concept of pre-trained weights from transfer learning and implementing a freeze training strategy,significant improvements have been made to enhance both the training speed and model training accuracy of this deep learning model.Finally,based on segmented blade surface defect images,a method for quantifying blade defects is proposed.This method combines image stitching algorithms to achieve overall quantification and risk assessment of the entire blade.Test results show that the improved Deeplabv3+deep learning model reduces training time by approximately 43.03%compared to the original model,while achieving mAP and MIoU values of 96.87%and 96.93%,respectively.Moreover,it demonstrates robustness in detecting different surface defects on blades across different back-grounds.The application of a blade surface defect quantification method enables the precise quantification of dif-ferent defects and facilitates the assessment of risk levels associated with defect measurements across the entire blade.This method enables non-contact,long-distance,high-precision detection and quantification of surface defects on the blades,providing a reference for assessing surface defects on wind turbine blades.

Numerical and Experimental Study on Flow-induced Noise at Blade-passing Frequency in Centrifugal Pumps

With the increasing noise pollution, low noise optimization of centrifugal pimps has become a hot topic. However, experimental study on this problem is unacceptable for industrial applications due to unsustainable cost. A hybrid method that couples computational fluid dynamics （CFD） with computational aeroacoustic software is used to predict the flow-induced noise of pumps in order to minimize the noise of centrifugal pumps in actual projects. Under Langthjem＇s assumption that the blade surface pressure is the main flow-induced acoustic source in centrifugal pumps, the blade surface pressure pulsation is considered in terms of the acoustical sources and simulated using CFX software. The pressure pulsation and noise distribution in the near-cutoff region are examined for the blade-passing frequency （BPF） noise, and the sound pressure level （SPL） reached peaks near the cutoff that corresponded with the pressure pulsation in this region. An experiment is performed to validate this prediction. Four hydrophones are fixed to the inlet and outlet ports of the test pump to measure the flow-induced noise from the four-port model. The simulation results for the noise are analyzed and compared with the experimental results. The variation in the calculated noise with changes in the flow agreed well with the experimental results. When the flow rate was increased, the SPL first decreased and reached the minimum near the best efficient point （BEP）; it then increased when the flow rate was further increased. The numerical and experimental results confirmed that the BPF noise generated by a blade-rotating dipole roughly reflects the acoustic features of centrifugal pumps. The noise simulation method in current study has a good feasibility and suitability, which could be adopted in engineering design to predict and optimize the hydroacoustic behavior of centrifugal pumps.

Collisionless tool orientation smoothing above blade stream surface using NURBS envelope

In five-axis machining,tool orientation above a blade stream surface may lead to tool collision and a decrease in workpiece rigidity.Hence,collisionless tool orientation smoothing(TOS)becomes an important issue.On the basis of a constant scallop height tool path,the triangular facets in the faces,vertices format are constructed from cutter contact(CC)using the Voronoi incremental algorithm.The cutter location(CL)points candidate set is represented by an oblique elliptic cone whose vertex lies at CC using NURBS envelope.Whether the CL point is above its CC is judged by the dot product between the normal vector and the point on triangulation nearest to the CL point.The curvatures at CC are obtained by fitting a moving least square(MLS) quadratic patch to the local neighborhood of a vertex and calculating eigenvectors and eigenvalues of the Hessian matrix.Triangular surface elastic energy is employed as the weight in selection from the NURBS envelope.The collision is judged by NURBS surface intersection.TOS can then be expressed by selecting a CL point for each CC point and converted into a numerical control(NC)code automatically according to the postprocessor type of the machine center.The proposed method is verified by finishing of a cryogenic turboexpander impeller of air separation equipment.

Surface integrity of GH4169 affected by cantilever finish grinding and the application in aero-engine blades

GH4169 is the main material for aero-cngine blades and integrated blisks. Because GH4169 has a poor milling performance, the profile precision and surtace integrity of blades and integrated blisks are difficult to be met by utilizing the conventional milling process, which directly influence the global performance and reliability of aero-engines. Through grinding experiments on parameters and surface integrity optimization, the helical cantilever grinding process utilizing a 300# CBN RB wheel is presented and applied in finish machining of GH4169 blades. The profile errors of the blade surface are within ±0.01 mm, the roughness is less than 0.4μm, the residual compressive stresses and the hardening rate are appropriate, there are no phenomena of burr and smearing with the grinding chips, and the leading/trailing edge can be smoothly connected with the suction/pressure surface. All the experimental results indicate that this grinding process is greatly suitable for the profile finish machining of GH4169 blades.

Experimental Study of Endwall and Tip Clearance Flows in a Two- Dimensional Turbine Rotor Blade Cascade - Effect of Incidence Angle

Experimental investigations were carried out on a two-dimensional cascade fitted with a 120° deflection rotor blades to study the effect of incidence angle on the endwall flow in the presence of tip clearance. A total of five incidence angles, namely: -10°，-10° -5°, 0°, 5° ， 10° were chosen and for each incidence angle, the experiments were conducted for five tip clearance values at a constant space -chord ratio of 0.79. The results are presented in the form of contours of static pressure coefficient on the endwall and the blade tip surface. In addition, the variation of static pressure coefficient ahead of the blade leading edge and from the pressure surface to the suction surface at various axial stations, and discharge coefficient gi different axial stations are presented. The results indicate that the adverse pressure gradient upstream of the leading edge is reduced as tip clearance is increased. The contours of static pressure coefficient on the endwall indicate a deep low-pressure trough near the suction surface in comparison to the normal trough for zero clearance. Loading also increases as incidence changes from the negative to positive values. Due to area contraction caused by the tip separation vortex, the fluid moving towards the tip gap from the pressure side is accelerated. Downstream of the tip separation vortex, the endwall pressure increases due to flow mixing. The maximum value of discharge coefficient increases and the point at which maximum value occurs shifts towards leading edge when incidence is changed from -10° to 10°.