基于特征参数的珩磨油石寿命预测研究

发布时间：2018-05-08 14:05

本文选题：珩磨机 + 油石磨损量　；参考：《兰州理工大学》2017年硕士论文

【摘要】：珩磨油石的磨损状态对产品的最终质量有着较大的影响。为了预测油石的切削寿命,便于合理的更换油石,通过对比油石磨损量与磨钝标准来判断油石是否需要更换。所以本文引入灰色神经网络,通过将珩磨工艺加工特征参数作为模型输入来预测油石的磨损量,最终建立了珩磨油石磨损量预报模型来预测油石的寿命。从而为提前更换油石提供了理论依据,在保证机床稳定运行,提高加工产品质量,节约制造执行系统中生产成本等方面具有重大意义。本论文主要研究内容包括:(1)研究了神经网络和灰色神经网络预报模型算法。由于神经网络具有高度非线性拟合能力以及珩磨加工本身可看做一个灰色系统,通过分析比对各种预报模型,最终选用以上两种模型。并对模型结构的选择,关键参数的设置进行了详细的阐述。给出了评价模型拟合精度和稳定性的依据。(2)研究了智能算法在预报模型优化中的运用。对比了粒子群算法(PSO),遗传算法(GA)以及蚁群算法(ACO)的优缺点。由于PSO算法具有收敛速度快,需要调整的参数较少等优点,采用该算法对模型进行优化。并根据算法存在的不足,提出了利用变异因子来对标准PSO算法进行优化改进,并利用目标函数,对算法的寻优能力和收敛性进行比较。(3)研究了适合珩磨油石磨损量预报的预报模型。以强力珩磨的数据为基础,建立了基于BPNN的珩磨油石磨损量预报模型,并利用MPSO算法和GA算法对其进行优化。由于珩磨加工可看为灰色系统,首先,利用灰色关联度分析了珩磨加工特征参数对珩磨油石磨损量的影响;其次,建立了基于GNN的油石磨损量组合预报模型,并利用MPSO算法对模型中的灰参数进行优化。通过仿真实验对比建立的各种模型,基于MPSO-GNN模型的MPAE值更小,说明该模型的精度更高,预测更稳定。因此,该模型在珩磨油石磨损量预测中具有一定的优势,可以用于实际加工中预测油石的磨损状态,进而合理更换油石。
[Abstract]:The wear state of honing stone has a great influence on the final quality of the product. In order to predict the cutting life of the oil stone and make it convenient to replace the oil stone, it is necessary to judge whether the oil stone needs to be replaced by contrasting the wear quantity and the bluntness standard of the oil stone. In this paper, grey neural network is introduced to predict the wear rate of honing stone by using honing process characteristic parameters as model input. Finally, the prediction model of honing stone wear quantity is established to predict the life of honing stone. Thus it provides a theoretical basis for the early replacement of oilstones, which is of great significance in ensuring the stable operation of machine tools, improving the quality of processed products, and saving the production cost in the manufacturing execution system. The main contents of this paper include: 1) the neural network and grey neural network prediction model algorithms are studied. Because the neural network has the ability of highly nonlinear fitting and honing itself can be regarded as a grey system, through the analysis and comparison of various prediction models, the two models are finally selected. The selection of model structure and the setting of key parameters are described in detail. The application of intelligent algorithm in prediction model optimization is studied. The advantages and disadvantages of particle swarm optimization (PSO), genetic algorithm (GA) and ant colony algorithm (ACO) are compared. Because the PSO algorithm has the advantages of fast convergence and few parameters to be adjusted, this algorithm is used to optimize the model. According to the shortcomings of the algorithm, this paper proposes to optimize and improve the standard PSO algorithm by using the mutation factor, and uses the objective function. The prediction model suitable for prediction of honing stone wear is studied by comparing the optimization ability and convergence of the algorithm. Based on the data of strong honing, the prediction model of honing stone wear quantity based on BPNN is established, and the MPSO algorithm and GA algorithm are used to optimize the model. Because honing can be seen as a grey system, firstly, the influence of honing characteristic parameters on honing stone wear is analyzed by grey correlation degree, secondly, the combined prediction model of honing stone wear quantity based on GNN is established. The grey parameters in the model are optimized by MPSO algorithm. Through the comparison of various models established by simulation experiments, the MPAE value based on MPSO-GNN model is smaller, which shows that the accuracy of the model is higher and the prediction is more stable. Therefore, the model has some advantages in predicting the wear volume of honing stone, which can be used to predict the wear state of the stone in practical processing and to replace the oil stone reasonably.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG580.67

【参考文献】