Surface roughness is influenced by various factors with uncertainty characteristics, and roughness reliability can be used for the assessment of the surface quality of CNC milling. The paper develops a method for the assessment of surface quality by considering the coupling effect and uncertainty characteristic of various factors. According to the milling kinematics theory, the milling surface topography simulation was conducted by discretizing the cutting edge, machining time, and workpiece. Considering the coupling effect of various factors, a roughness prediction model is established by the SSA-LSSVM, and its prediction accuracy reaches more than 95%. Then, the roughness reliability model was developed by applying the response surface methodology to achieve the assessment of surface quality. The proposed method is verified by the milling experiments. The maximum values of the relative errors between the simulation and experimental results of the surface roughness and roughness reliability are 9% and 1.5% respectively, indicating the correctness of the method proposed in the paper.
Since the emergence of artificial intelligence and deep learning methods, the fault diagnosis of bearings in rotating machinery has gradually been realized, reducing the high costs of bearing faults. However, in the actual work of the equipment, faults rarely occur, resulting in less fault data. Therefore, it is necessary to study small sample fault data. For the case of less fault data, the Maml–Triplet fault classification learning framework based on the combination of maml and the triplet neural network is proposed. In the framework of Maml-Triplet fault classification, firstly, an initial signal feature extractor is obtained using the Maml training method. Secondly, the feature vectors corresponding to signal data are obtained using depth distance measurement learning in the triplet neural network, and the fault type is judged based on the feature vectors of unknown signal. The results show that the accuracy of the Maml–Triplet model is 2% higher than that of the triplet model alone and 5% higher than that of the Maml–CNN meta learning method. When there are fewer data samples, the accuracy gap is more obvious. Therefore, in the case of less data, the Maml–Triplet model has an excellent fault identification ability.
Four-axis machine tools with two rotary axes are widely used in the machining of complex parts. However, due to an irregular kinematic relationship and non-linear kinematic function with geometric error, it is difficult to analyze the influence the geometry error of each axis has and to compensate for such a geometry error. In this study, an influence analysis method of geometric error based on the homogeneous coordinate transformation matrix and a compensation method was developed, using the Newton iterative method. Geometric errors are characterized by a homogeneous coordinate transformation matrix in the proposed method, and an error matrix is integrated into the kinematic model of the four-axis machine tool as a means of studying the influence the geometric error of each axis has on the tool path. Based on the kinematic model of the four-axis machine tool considering the geometric error, a comprehensive geometric error compensation calculation model based on the Newton iteration was then constructed for calculating the tool path as a means of compensating for the geometric error. Ultimately, the four-axis machine tool with a curve tool path for an off-axis optical lens was chosen for verification of the proposed method. The results showed that the proposed method can significantly improve the machining accuracy.
10 July 2026 in Eksploatacja i Niezawodność – Maintenance and Reliability
Demands for superior surface quality in high-end components make precise analysis and reliable assessment critical. This paper presents an integrated method combining improved morphology simulation with an intelligent evaluation framework. Specifically, the Z-Map algorithm is enhanced by particle swarm optimization for efficient, high-precision surface topography simulation. The high-fidelity simulation data provides an effective database for training the subsequent intelligent prediction model. The simulated data then trains a radial basis function neural network (RBFNN) optimized by an improved butterfly optimization algorithm (IBOA) for roughness prediction, achieving over 99% accuracy. This model serves as a surrogate in Monte Carlo simulations for reliability analysis (IBOA-RBFNN-MC). Results show this approach outperforms benchmarks and aligns closely with experimental measurements, validating its accuracy for process optimization and quality control.