液压支架掩护梁结构优化及疲劳寿命预测

Structure Optimization on Shield Beam of Hydraulic Support and Prediction on Its Fatigue Life

针对传统的液压支架设计具有耗时、耗材、成本高等缺陷,提出一种利用ANSYS和BP神经网络相结合对掩护梁进行轻量化和寿命预测的设计方法。首先应用ANSYS软件参数化语言APDL对ZY18000-25-45型液压支架掩护梁进行建模;再应用ANSYS/OPT模块对掩护梁进行尺寸优化分析,使其在满足液压支架设计准则的前提下,对掩护梁进行轻量化设计,并对原有掩护梁模型进行改造,整体质量减少了6.9%;最后利用遗传算法改进BP神经网络对优化后的掩护梁建立神经网络模型,对掩护梁的疲劳寿命进行预测。结果表明,掩护梁平均寿命为17228次,优化后的掩护梁满足可靠性要求。

The traditional design of hydraulic support had the disadvantages of time consuming, consumables and high cost. In view of these, a new design method combining AN- SYS with BP neural network was used for the lightweight and life prediction of shield beam. Firstly, the model of shield beam for ZY18000- 25- 45 type hydraulic support was established by parametric design language（APDL） of ANSYS. And the size of shield beam was optimized and ana- lyzed to meet the design criteria by ANSYS/OPT module. On this basis, the whole quality was decreased by 6.9% after lightweight and reconstruction for original hydraulic support. Finally, the BP neural network improved by genetic algo- rithm was adopted to build the neural network model for op- timized shield beam, and to forecast its fatigue life. The re-suits showed that the average life of shield beam was 17228 times, which met the reliability requirement after optimization.